自動駕駛模擬及世界模型研究報告（2026）

自動駕駛模擬研究－「模擬測試+世界模型」主導的測試系統構成了研發的基礎。

「自動駕駛模擬與世界模型研究報告（2026）」主要聚焦於模擬與世界模型領域的核心技術、產業趨勢與主流解決方案。報告涵蓋了完整的模擬測試體系（從MIL到VIL的X-in-the-loop測試、場景庫建置等），以及OEM廠商和一級供應商的世界模型解決方案演進。報告分析了國內14家領先的模擬平台和世界模型解決方案供應商以及海外13家，總結了模擬測試與世界模型之間的協同效應。透過這項研究，報告論證了世界模型在降低資料成本、泛化情境以及輔助決策推理方面的核心價值。

國家標準「GB/T 47025-2026 智慧網聯汽車－自動駕駛功能模擬測試方法及要求」於2026年1月28日發布並正式實施。此標準適用於配備自動駕駛功能或系統的M類和N類車輛，規定了自動駕駛功能的模擬試驗方法、試驗要求和整體評估標準。該標準共定義了七大類48個專項試驗項目。 GB/T 47025-2026與已發布的GB/T 41798-2022「智慧網聯汽車－自動駕駛功能場地試驗方法及要求」和GB/T 44719-2024「智慧網聯汽車－自動駕駛功能公共道路測試方法及「要求」共同構成了三系統採用完整的檢驗系統，該系統採用完整的「公用道路配置」與「公共道路配置」。

為了加速L3/L4級自動駕駛車輛的量產，成熟的自動駕駛演算法檢驗通常採用「99.9%模擬測試+0.09%封閉場地測試+0.01%公共道路測試」的黃金比例。建議的國家標準GB/T 47025-2026要求感測器模型與實際車輛的誤差≤5%；動力學模型與實際車輛的一致性≥95%；並且交通參與者的行為具有高度逼真性。這標誌著自動駕駛產業已進入以合規性和安全性為優先發展的新階段。模擬測試不再只是研發的輔助工具，而是產品市場進入、認證和安全驗證中一項具有法律強制性的流程。

另一方面，作為生成式人工智慧模型的世界模型，能夠透過建構內部表徵來理解現實世界的動態規律（包括物理屬性和空間屬性）。它們還能根據文字、圖像、影片和運動數據等輸入資訊產生影片內容。這項技術在自動駕駛和機器人等領域展現出巨大的應用潛力，並正成為推動智慧系統提升感知和決策能力的核心技術支柱。

1. 模擬平台正在發展成為具有高保真度、物理一致性和動態互動的「訓練環境」。

模擬系統的角色已從傳統的測試執行工具提升為支援演算法訓練的核心資料基礎設施。此外，仿真環境也從視覺相似性發展到行為真實性，旨在透過強調對物理感測器（光子、電訊號、多回波等）、精確的材料屬性（反射率、粗糙度等）以及符合物理定律的車輛動力學和交通流的仿真，來彌合「仿真到現實」之間的差距。

在高保真度方面，模擬平台供應商不斷提升模擬檢驗能力，讓可靠的模擬更加精細。例如，Keymotek 的 aiSim6 符合 ASAM OpenMATERIAL 3D 標準，可透過定義精確的材質物理屬性，提供諸如相機（非線性響應、CMOS 雜訊）和雷射雷達（高斯光線、多回波、天氣衰減）等物理感測器的模擬。此外，基於其專有的 PBR 噴濺技術，aiSim6 可以動態調整 3DGS 模型的場景光照。這使得在同一路段內動態切換白天、黃昏和夜晚等光照條件成為可能，從而將其轉化為「動態可配置的訓練環境」，並實現「物理動態神經渲染」。

具體來說，aiSim 6說明流體運動的納維-斯托克斯方程式應用於環境粒子物理模擬，從而將物理環境擾動引入合成資料鏈路。這使得模擬諸如車輛氣流引起的樹葉運動、雨天路面濺起的水花以及井蓋蒸氣與交通參與者之間的動態相互作用等現象成為可能，克服了物理真實性在極端場景下的不足。

在物理一致性方面，以PilotD Technology的高保真物理模擬為例。該公司自主研發了自演化雙渦輪驅動的資料訓練平台。該平台利用高保真世界模型產生視覺、點雲等多模態數據，用於機器人大腦的封閉回路型訓練。同時，該公司的資料可靠性檢驗技術「物理評判」系統，對產生資料的物理有效性進行檢驗，並同時進行資料篩檢和世界模型的封閉回路型訓練。基於這個自演化數據雙渦輪，EAI的「Cerebrum」透過注入具有更高物理有效性的合成數據，實現全自動迭代演化，從而提升演算法在複雜真實場景中的適應性和泛化能力。

該公司專有的全物理光學核心建模技術能夠高度精確地再現資料的光學物理特性。此技術用於訓練多模態世界模型資料生成架構，在動態特性和光學特性方面均具有高保真度，從而為人工智慧公司提供高保真度的合成資料解決方案。

例如，在動態互動方面，SYNKROTRON 的 OASIS Traffic 解決方案是一個高階自動駕駛的交通流合成資料平台，它基於實際路邊資料。該方案利用人工智慧生成涵蓋 60 種高階互動場景的對抗性交通流，使用 TTC/PET 量化風險，並涵蓋超過 30% 的長尾特殊案例。它還可以產生大規模動態交通場景資料集（典型區域、典型交通場景、動態參與者、自然行為和對抗行為）。

2. 世界模式將從輔助工具轉變為核心基礎。

世界模型旨在透過內化重力、碰撞和因果關係等物理定律來理解世界的「常識」，從而解決傳統模擬工具在長期一致性和可解釋性方面存在的問題。例如，GigaAI 的 GigaWorld-1 對物理定律的遵循度極高，能夠精確模擬重力和碰撞等複雜的物理交互作用。 Li Auto 的 MindVLA-o1 利用原生 3D ViT 和預測性潛在世界模型來理解3D空間結構中物件的位置關係和運動模式。此外，世界模型還可用於產生大量、高保真且多樣化的訓練數據，從而解決真實世界物理交互數據極度匱乏的問題，並促進「模擬到現實」的轉變。

融合趨勢：VLA + 世界模型 + 強化學習

在自動駕駛領域，世界模型已從單純的資料產生工具發展成為自動駕駛系統認知和推理的核心，並與虛擬語言分析（VLA）和強化學習深度整合。在演算法訓練中，VLA負責感知和語意理解，世界模型負責未來推理和預測，而強化學習則負責虛擬世界中的自主最佳化決策。這三者協同工作。例如：

QCraft的「VLA+世界模型」整合架構不僅整合了經數百萬台量產驗證的端對端功能，而且透過語言處理能力，能夠精準理解環境文本、複雜場景和語音指令，從而協調模型決策、遠端控制和人機交互三大要素。此外，借助世界預測模型，它還能準確推斷交通參與者的行為、道路結構的變化以及動態場景的發展，進而規劃出最優的行車軌跡。

作為VLA 2.0的「雲矩陣」，小鵬X-World是一款能夠「思考」駕駛場景的實體AI模擬器。它透過世界模型產生大量場景，進行訓練和評估，並將研發範式從「累積真實世界測試」轉變為「累積運算能力訓練」。模型基於先進的影像生成模型WAN 2.2，並採用客製化的DiT骨幹網路。其主要創新之處在於引入了觀點-時間自注意力機制。此機制迫使模型在生成過程中同時對時間軸和周圍七個攝影機觀點之間的空間幾何關係進行建模，從而確保生成的虛擬世界在觀點之間緊密銜接，防止物體「穿過模型」或錯位。底層採用高壓縮性的3D因果變分自編碼器（VAE），顯著降低了多通道影片串流處理的計算量，並支援長期建模。

世界模型的關鍵應用：

在自動駕駛領域，世界模式採用「雲端學習+車載推理」的雙引擎架構。雲端負責大規模訓練和場景生成，而車輛則進行即時決策並快速回應。例如，華為於2026年4月24日發布了乾坤ADS 5，該系統採用WEWA 2.0演算法，基於博弈論將訓練和學習效率提升10倍，碰撞風險降低50%。 2026年，雲端運算算力將大幅提升至目前的60 EFLOPS，比2023年成長21倍，為先進的自動駕駛研發提供強力支撐。

在華為的WEWA架構中，基於雲端的WE（世界引擎）負責虛擬場景的訓練和模型參數的更新。它基於擴散生成模型，以同步生成、學習和檢驗的模式運作。它可以可控地產生各種罕見場景，例如相鄰車輛切入、車輛突然駛出和領先緊急制動，從而實現從人工輔助AI訓練到AI自主訓練的轉變。車聯網WA（世界動作模型）則負責即時路徑規劃與控制。

作為世界級模型，Pony.ai 的 PonyWorld 2.0 具備自我診斷和定向演化能力。此人工智慧能夠獨立診斷問題並主動引導資料收集，代表研發訓練模式轉移。具體而言，PonyWorld 2.0 與 Pony.ai 汽車模型的意圖語意層結合，能夠對所有駕駛決策進行自動追溯和歸因分析。該系統能夠自動識別問題的根本原因，並將診斷結果準確反映在模型訓練過程中。

基於自我評估結果，PonyWorld 2.0 可自動辨識世界模型精確度不足的特定場景，並主動產生定向資料擷取任務。例如，系統可以自動推送指令，例如「在特定時間內，重點收集指定路口逆光條件下，非機動車和行人混合交通場景的資料」。這使得研發和測試團隊能夠圍繞世界模型的「精度要求」高效協作，從而實現人工智慧引導的定向資料收集和模型迭代。

在企業人工智慧 (EAI) 領域，世界模型已經從「數據引擎」演變為能夠進行物理推理、行動規劃和任務確定的 EAI 代理的「大腦」或「模擬器」。

例如，與依賴低效且耗時的影像預測連結的傳統 WA 架構不同，GigaAI 以動作為中心的範式，即 GigaWorld-Policy 的世界動作模型 (WAM)，消除了跨模態耦合瓶頸，並透過架構最佳化顯著提高了推理效率。

這開創了一種「複雜訓練與簡化推理」的混合範式模型：

在學習階段，GigaWorld-Policy 使用因果掩蔽機制來實現動作標記和未來視覺標記的統一模型，從而使動作預測能夠充分利用未來視覺動態提供的高密度監督訊號。

在推理階段，該模型完全消除了影片預測分支，僅保留了一個輕量級的動作生成模組。這避免了對長視覺標記序列進行推理處理，並從根本上消除了傳統WA模型中跨模態架構耦合導致的結構性計算冗餘。

與目前主流的廣域網路模型（例如 Motus 和 Cosmos Policy）相比，GigaWorld-Policy 在維持策略品質的前提下，推理速度提升了 10 倍，真正滿足了機器人高頻封閉回路型控制的即時性要求。 GigaWorld-Policy 在實際任務中的平均成功率接近 85%。即使與 Cosmos-Policy 等強大的競爭模型相比，其絕對成功率也高出 30% 以上。

2026年4月29日，GensPi Technology正式發表了通用世界動作模型MotuBrain。 MotuBrain定位為EAI機器人的通用大腦，具備多機器人適應性、多任務泛化能力和長期任務執行能力，實現了單腦多功能性和多類型相容性。 MotuBrain的核心創新在於其對「感知世界」和「待執行動作」的統一建模，使機器人不僅能夠理解環境，還能預測環境變化並產生可執行的動作策略。 MotuBrain在RoboTwin 2.0和WorldArena兩大國際權威評測中均榮獲第一名。在WorldArena評測中，MotuBrain以63.77的綜合EWM得分位列第一，並在運動品質、流暢度、運動平滑度等多個關鍵運動指標中均表現出色。

3. 「模擬測試+世界模型」主導的測試體系成為研發的基礎。

在自動駕駛的封閉回路型數據和測試系統中，模擬測試和世界模型相互補充，抵消技術上的不足，彌補彼此能力上的限制。

在自動駕駛和企業應用人工智慧（EAI）領域，模擬測試和世界模型正從「分離」轉向「深度融合」。業界正著手建立統一標準，並推動建構融合「重構+生成+模擬+訓練」的整合平台，旨在實現從自動駕駛到企業應用人工智慧的模擬能力，並建構更廣泛的實體人工智慧生態系統。

目前，世界模型（尤其是生成式世界模型）是模擬平台的核心「動力來源」，驅動人工智慧自動生成模擬場景。這使得低成本、高品質地產生海量且多樣化的場景（尤其是長尾和罕見場景）以及高保真度的感測器資料成為可能。

2026年4月24日，51Sim的SimOne 4.0針對物理AI時代進行了全面重構和升級，建構了「4DGS重建+生成式世界建模」的技術基礎。這使得系統能夠基於真實車輛資料自動建立互動式、可編輯、可擴展的虛擬模擬資產，實現大規模場景生成。 SimOne 4.0涵蓋資料、訓練、推理、檢驗和交付五大核心流程，全面支援AI安全且有效率地進入實體世界。此外，SimOne 4.0在產品層面深度整合了NVIDIA Omniverse NuRec神經渲染技術解決方案，打造了從真實世界資料擷取、神經場景重建到封閉回路型執行的完整資料驅動流程。在 51Sim 的端對端資料驅動封閉回路型解決方案中，動力學、雷射雷達和相機模擬的置信度分別達到 95%、95% 和 90%，模擬測試與現場測試的一致性達到 92%。

SimOne 4.0 支援多種 GPU 架構的同步運作。它針對 Moore Threads 的旗艦級 AI 訓練推理整合 GPU——MTT S5000 進行了系統化的適應和全面的最佳化。該平台能夠高度並行地執行大規模4DGS 和世界模型訓練任務，在短時間內為複雜的動態場景提供高品質的重建和模型訓練，並推動世界模型和 VLA 的持續演進。迄今為止，SimOne 已為自動駕駛、智慧型裝置和機器人等眾多 EAI 領域的 100 多家客戶提供支援。

2026年1月，AGIBOT發布了基於其大規模語言模型的開放原始碼平台Genie Sim 3.0。該平台基於NVIDIA Isaac Sim構建，提供高保真模擬環境和自然語言場景生成功能。它提供了閉合迴路解決方案，涵蓋從數位資產生成、場景泛化到資料收集和自動化評估的整個流程，顯著加快了模型訓練和檢驗過程，同時降低了對實體硬體的依賴。

Genie Sim 3.0 的主要特性包括數位雙胞胎等級的高保真模擬環境。該環境突破性地整合了3D重建、視覺生成技術和實體引擎，實現了統一的視覺真實感和物理精確度。其次，它率先實現了自然語言驅動的場景生成和泛化。透過 Genie Sim 3.0，開發者只需輸入自然語言指令，平台即可在幾分鐘內自動產生並泛化數千個訓練和測試場景，從而實現大規模並行訓練。此外，此模擬平台還提供了一個全面的開放原始碼模擬資料集（涵蓋 200 多個任務，累計數萬小時）和高效的資料集擷取方案。它還建構了3D評估系統。值得一提的是，AGIBOT 的世界模式「Genie Envisioner」是基於 NVIDIA Cosmos，實現了從感知到行動的端到端封閉回路型。 GE 將政策學習、評估和模擬功能整合到一個統一的框架中，該框架以統一的影片生成世界模型為核心。 AGIBOT 透過將 Cosmos Predict 2 深度整合到其專有的行為條件世界模型架構中，為 GE-Sim 提供強大的通用視覺和物理先驗知識功能。

模擬測試與世界模型的融合，本質上形成了一個「資料生成—演算法學習—模型檢驗—持續演化」的飛輪式封閉回路型。在自動駕駛和企業人工智慧（EAI）這兩個領域，這種融合路徑高度一致，二者都指向「物理人工智慧」的最終目標。這意味著系統能夠在虛擬世界中完成從認知到行動的封閉回路型學習，並無縫過渡到物理世界。

目錄

條款

第1章：仿真測試概述

• 模擬測試階段及協作
  • 仿真場景建模技術
  • 3DGS高精度模擬場景重建技術
  • 仿真場景建模
  • 3DGS技術的四大發展方向
  • 用於自動駕駛場景的4D重建和動態建模的通用演算法
  • 仿真場景建模
  • 物理感測器仿真
  • 透鏡模型的光學和物理建模
  • CMOS感測器的光電子模擬
  • LiDAR建模：高斯射線和物理衰減
  • 案例研究：LidarPainter
  • 車輛動力學模擬解決方案比較
  • 案例研究：Panocar 2026 新版本
  • X 在環測試和調整
  • SIL案例研究
• 模擬場景庫
  • 場景抽象層次結構
  • 仿真場景庫的分類
  • 場景庫資料來源
• OpenX 系列標準
• OpenX系列標準在仿真平台中的應用範例
• 交通流合成資料平台範例
• 交通流合成資料集範例：
• 非結構化道路模擬面臨的挑戰
• 非結構化道路模擬中克服困難的案例研究
• 模擬測試與世界模型之間的關係
• 「模擬測試+世界模型」驅動測試系統
• 模擬測試/世界模型的發展趨勢 1
• 國家標準GB/T 47025-2026「智慧網路聯網汽車－自動駕駛功能模擬試驗方法及要求」的解釋
• 國家模擬標準中48項測試項目清單
• 模擬測試與全球模型產業鏈

Autonomous driving simulation research: "Simulation test + world model"-driven test system has become R&D infrastructure.

The "Autonomous Driving Simulation and World Model Research Report, 2026" mainly focuses on core technologies, industry trends and mainstream solutions in the field of simulation and world models, covering the complete system of simulation testing (X-in-the-loop testing from MIL to VIL, scenario library construction, etc.), as well as the evolution of world model solutions of OEMs/Tier 1 suppliers. It analyzes 14 Chinese and 13 foreign mainstream simulation platforms and world model solution providers, sorts out the synergistic relationship between simulation testing and world models, and demonstrates the core value of world models in data cost reduction, scenario generalization, and decision reasoning by way of research.

The national standard GB/T 47025-2026 Intelligent and Connected Vehicle - Simulation Test Methods and Requirements for Automated Driving Function was released and officially implemented on January 28, 2026. This standard is applicable to Category M and N vehicles with autonomous driving functions or autonomous driving systems, and stipulates the simulation test methods, test requirements and overall criteria for the autonomous driving function. The standard defines a total of 48 special test items in 7 categories. The GB/T 47025-2026 and the released standards, the GB/T 41798-2022 (Intelligent and Connected Vehicles - Track Testing Methods and Requirements for Automated Driving Function) and the GB/T 44719-2024 (Intelligent and Connected Vehicle - Methods and Requirements of Road Test for Automated Driving Functions), constitute a complete verification system of "simulation-field-road" trinity.

In order to speed up the mass production of L3/L4 autonomous vehicles, mature autonomous driving algorithm verification usually follows the golden ratio of "99.9% simulation tests + 0.09% closed field tests + 0.01% public road tests". The recommended national standard GB/T 47025-2026 requires that the error between the sensor model and the actual vehicle should be <=5%; the consistency between the dynamics model and the actual vehicle should be >=95%; the behavior of traffic participants should be high-fidelity, etc. This means that the autonomous driving industry has entered a new development stage of compliance access and safety priority. Simulation testing is no longer an auxiliary means for research and development, but has become a legally required link for product access, certification, and safety evidence.

Meanwhile, as a generative AI model, the world model can understand the dynamic laws of the real world (covering physical characteristics and spatial attributes) by building internal representations. It also generates video content with input information such as text, images, videos, and motion data. It is quickly showing great application potential in fields such as autonomous driving and robotics, and is becoming the core technical pillar that drives intelligent systems to leap into high-level perception and decision capabilities.

1. The simulation platform evolves into a "training environment" with high fidelity, physical consistency, and dynamic interaction.

The positioning of the simulation system has been upgraded from a traditional test execution tool to a core data infrastructure supporting algorithm training. The simulation environment is also evolving from visual resemblance to behavioral authenticity, emphasizing physical sensor simulation (such as photons, electrical signals, multi-echo), accurate material properties (such as reflectivity, roughness), vehicle dynamics and traffic flow that comply with physical laws, in a bid to bridge the "Sim-to-Real" gap.

In terms of high fidelity, simulation platform companies are continuing to upgrade their simulation verification capabilities, making high-confidence simulations more detailed. For example, Keymotek's aiSim6 can provide physical sensor simulations, such as cameras (nonlinear response, CMOS noise) and LiDAR (Gaussian rays, multi-echo, weather attenuation), following the ASAM OpenMATERIAL 3D standard and defining precise material physical properties. Furthermore, based on its self-developed PBR Splatting technology, it can dynamically adjust scenario lighting for 3DGS models, dynamically switching lighting conditions such as daytime, dusk, and nighttime on the same road segment, transforming it into a "dynamically configurable training environment" and achieving "physical dynamic neural rendering."

Notably, aiSim 6 applies the Navier-Stokes equations describing fluid motion to environmental particle physics simulations, introducing physical environmental disturbances into the synthetic data link. This allows for realistic simulations of leaf movement caused by vehicle airflow, water splashes from pavements during rain, and the dynamic interaction between manhole cover steam and traffic participants, addressing the shortcomings in physical realism of edge scenarios.

In terms of physical consistency, take the high-fidelity physical simulation of PilotD Technology as an example. The company has independently developed a self-evolving dual-turbine driven data training platform. It uses a high-fidelity world model to generate multi-modal data such as vision and point clouds for closed-loop training of the robot brain. Meanwhile, its data credibility verification technology, namely the "Physical Judge" system, checks the physical rationality of generated data, and performs data screening as well as closed-loop retraining of the world model simultaneously. Based on the self-evolving data dual-turbine, the EAI cerebrum completes fully automatic iterative evolution with the injection of increasingly physically relevant synthetic data, enhancing the algorithm's adaptability and generalization capability in complex real-world scenarios.

The company's self-developed fully physical optical core modeling technology highly restores the optical physicality of data, and uses this to train a multimodal world model data generation architecture with high fidelity in both dynamics and optics, providing AI companies with high-fidelity synthetic data solutions.

In terms of dynamic interaction, for example, SYNKROTRON's OASIS Traffic solution, a traffic flow synthesis data platform for advanced autonomous driving, is based on real roadside data. It uses AI to generate adversarial traffic flows covering 60 high-interaction scenarios, quantifies hazard levels using TTC/PET, and covers over 30% of long-tail corner cases. It can generate massive dynamic traffic scenario datasets (typical areas, typical traffic scenarios, dynamic participants, natural and confrontational behaviors).

2. The world model transforms from an auxiliary tool to a core foundation.

The world model is committed to internalizing physical laws, such as gravity, collision, and causality, to solve problems with traditional simulation tools such as long-term consistency and interpretability, and to understand the "common sense" of the world. For example, GigaAI's GigaWorld-1 has excellent physics adherence capabilities and can accurately simulate complex physical interactions such as gravity and collision. Li Auto's MindVLA-o1 uses the native 3D ViT and the predictive latent world model to understand object position relationships and movement patterns in the three-dimensional space structure. It makes use of the world model to generate massive, high-fidelity, and diverse training data to handle the extreme scarcity of real physical interaction data and promote "Sim2Real" migration.

Fusion trend: VLA + world model + reinforcement learning

In the field of autonomous driving, the world model has been upgraded from a single data generator to the core cognition and deduction center of the autonomous driving system, deeply integrated with VLA and reinforcement learning. In algorithm training, VLA is responsible for perception and semantic understanding, the world model for future deduction and prediction, and reinforcement learning for autonomous optimization decision in the virtual world. The three work together. For example,

QCraft's "VLA + world model" unified architecture can not only multiplex end-to-end capabilities that have been verified in millions of mass productions, but can also accurately understand environmental text, complex scenarios and voice commands through language capabilities, achieving triple alignment of model decision, teleoperation and HMI; then with the help of the world prediction model, it can accurately deduce the behavior of traffic participants, road structure changes and dynamic scenario evolution, thereby planning the optimal driving trajectory.

As the "cloud matrix" of VLA 2.0, XPeng X-World is a physical AI simulator that can "think" about driving scenarios. It generates massive scenarios through the world model for training and evaluation, and enables the R&D paradigm to shift from "stacking real vehicle testing" to "stacking computing power training." The model is built based on the leading video generation model WAN 2.2, involving a customized DiT backbone network. Its key innovation lies in the introduction of a perspective-time self-attention mechanism, which forces the model to simultaneously model the temporal dimension and the spatial geometric relationship between the seven surround view camera perspectives during generation, thereby ensuring that the generated virtual world is tightly integrated across perspectives and avoiding objects from "crossing the model" or being misaligned. The underlying layer adopts a 3D causal variational autoencoder (VAE) with high compression ratio, which greatly reduces the computational overhead of multi-channel vide o stream processing and supports long-term modeling.

Core Foundation Cases of World Models:

In the field of autonomous driving, the world model adopts a dual-engine architecture of "cloud training + vehicle reasoning". The cloud is responsible for large-scale training and scenario generation, and the vehicle offers real-time decision and rapid response. For example, on April 24, 2026, Huawei released Qiankun ADS 5, which uses the WEWA 2.0 to improve game-theoretic training and learning efficiency by 10 times, and reduce collision risks by 50%; cloud computing power jumps to the current 60 EFLOPS in 2026, achieving a 21-fold increase from the level in 2023, supporting high-level autonomous driving research and development.

In Huawei's WEWA architecture, the cloud-based WE (World Engine) handles virtual scenario training and model parameter updates. Powered by diffusion generative models, it operates in a mode of simultaneous generation, learning and validation. It can controllably generate various rare scenarios including adjacent vehicle cut-in, dart-out, and sudden braking of leading vehicles, realizing the shift from human training AI to AI self-training. The automotive WA (World Action Model) is in charge of real-time path planning and control.

As a world model, Pony.ai's PonyWorld 2.0 has self-diagnosis and directional evolution capabilities. AI can independently diagnose shortcomings and proactively guide data collection, becoming the core of the paradigm shift in R&D training. Specifically, PonyWorld 2.0 combines the intention semantics layer of Pony.ai's automotive model to realize automated traceback and attribution analysis of every driving decision. The system can automatically identify the root cause of the problem and accurately feed the diagnosis results to the model training process.

Based on self-diagnosis results, PonyWorld 2.0 can automatically identify specific scenarios where the accuracy of the world model is insufficient, and proactively generate directional data collection tasks. For example, the system can automatically push instructions: "Please focus on collecting mixed traffic scenario data of non-motorized vehicles and pedestrians under backlight conditions at designated intersections during specific periods." The R&D and testing teams thus collaborate efficiently around the "accuracy requirements" of the world model to achieve directional data collection and model iteration guided by AI.

In the field of EAI, the world model has evolved from a "data engine" to a "cerebrum" or "simulator" of EAI agents, capable of physical deduction, action planning and mission decision.

For example, unlike the traditional WA architecture that relies on inefficient and lengthy video prediction links, the action-centric paradigm of GigaWorld-Policy, the World-Action Model (WAM) developed by GigaAI, breaks the cross-modal coupling bottleneck and delivers a dramatic improvement in inference efficiency via architectural optimization.

It has pioneered the hybrid paradigm model of "Complex Training & Simplified Inference":

During the training phase, GigaWorld-Policy uses a causal mask mechanism to achieve unified modeling of action tokens and future visual tokens, allowing action prediction to fully benefit from the high-density supervision signals provided by future visual dynamics.

During the inference phase, the model completely abandons the video prediction branch, retaining only a lightweight action generation module. It avoids the need to perform inference processes for long sequences of visual tokens, fundamentally circumventing the structural computational redundancy caused by cross-modal architecture coupling in traditional WA models.

Compared to current mainstream WA models (such as Motus and Cosmos Policy), GigaWorld-Policy achieves a 10x improvement in inference speed while maintaining policy quality, truly meeting the real-time requirements of high-frequency closed-loop control for robots. GigaWorld-Policy's average success rate in real-world tasks approaches 85%. Facing strong competitors like Cosmos-Policy, its absolute success rate is raised by more than 30%.

On April 29, 2026, GensPi Technology officially released MotuBrain, a general-purpose world-action model. Positioned as a general-purpose cerebrum for EAI robots, it possesses multi-robot adaptability, multi-task generalization, and long-term task execution capabilities, achieving multi-functionality and multi-type capabilities with a single brain. MotuBrain's core breakthrough lies in its unified modeling of the "world seen" and the "actions to be performed," allowing the robot to not only understand the environment but also predict changes and generate executable action strategies. MotuBrain won the first place on both RoboTwin 2.0 and WorldArena, two authoritative international benchmarks. In WorldArena, MotuBrain ranked first with an overall EWM score of 63.77, and led across multiple key motion dimensions, including Motion Quality, Flow Score, and Motion Smoothness.

3. "Simulation test + world model"-driven test system has become R&D infrastructure.

In the autonomous driving data closed-loop and test system, simulation testing and world models complement each other, offsetting technical shortcomings and complement capability boundaries of each other.

In the fields of autonomous driving and EAI, simulation testing and world models are moving from "separation" to "deep integration." The industry has begun to establish unified standards and promote the construction of an integrated platform of "reconstruction + generation + simulation + training" to enable simulation capabilities from autonomous driving multiplexing to EAI, realizing a broader physical AI ecosystem.

Currently, world models (especially generative world models) have become the core "power plant" of simulation platforms, driving the AI-powered automatic generation of simulation scenarios and generating massive and diverse scenarios (especially long-tail and rare scenarios) and high-fidelity sensor data at low cost and with high quality.

On April 24, 2026, 51Sim's SimOne 4.0 was comprehensively reconstructed and upgraded for the physical AI era, building a "4DGS reconstruction + generative world model" technology base to automatically build interactive, editable, and scalable virtual simulation assets from real vehicle data to achieve large-scale scenario generation. SimOne 4.0 covers the five-core links of data, training, reasoning, verification and delivery, comprehensively helping AI enter the physical world safely and efficiently. Moreover, SimOne4.0 deeply integrates the neural rendering technology solution - NVIDIA Omniverse NuRec at the product level to build a complete data-driven process from real data collection, neural scenario reconstruction to closed-loop simulation execution. In 51Sim's end-to-end data-driven closed-loop solution, the confidence levels of dynamics, LiDAR, and camera simulations are as high as 95%, 95%, and 90% respectively, and the consistency between simulation testing and field testing reaches 92%.

SimOne 4.0 supports multiple GPU architectures simultaneously. It has achieved systematic adaptation and in-depth optimization with Moore Threads' flagship AI training and inference integrated GPU MTT S5000. The platform enables high-concurrency execution of large-scale 4DGS and world model training tasks, delivering high-quality reconstruction and model training for complex dynamic scenarios within a short time, and driving the continuous evolution of world models and VLA. Up to now, SimOne has empowered more than 100 customers in many EAI fields such as autonomous driving, smart equipment, and robots.

In January 2026, AGIBOT released Genie Sim 3.0, an open-source simulation platform driven by its large language model. Based on NVIDIA Isaac Sim, the platform provides a high-fidelity simulation environment and natural language-driven scenario generation capabilities. It can provide a full-process closed-loop solution from digital asset generation, scenario generalization, data collection to automatic evaluation, significantly speeding up the model training and verification process and reducing dependence on physical hardware.

Highlights of Genie Sim 3.0 include a digital twin-level high-fidelity simulation environment, which pioneeringly deeply integrates three-dimensional reconstruction, visual generation technology and physics engines to achieve the unification of visual realism and physical accuracy. Secondly, it has pioneered natural language-driven scenario generation and generalization. In Genie Sim 3.0, developers can input natural language instructions to drive the platform to automatically generate and generalize thousands of training and test scenarios within minutes, and conduct large-scale parallel training. In addition, the simulation platform also provides a full range of open-source simulation datasets (covering more than 200 tasks with a total duration of tens of thousands of hours) and efficient collection solutions; it has built a three-dimensional evaluation system based on 100,000+ simulation scenarios, etc. It is worth noting that AGIBOT's world model, Genie Envisioner, is based on NVIDIA Cosmos to realize an end-to-end closed loop from perception to action. GE uses a unified video generative world model as the core to integrate policy learning, evaluation and simulation capabilities into the same framework. AGIBOT provides GE-Sim with powerful general visual and physical prior capabilities by deeply integrating Cosmos Predict 2 into its self-developed action-conditioned world model architecture.

The integration of simulation testing and world models essentially builds a flywheel closed loop of data generation - algorithm training - model verification - continuous evolution. In the two fields of autonomous driving and EAI, the integration paths are highly consistent, both pointing to the ultimate goal of "physical AI": allowing the system to complete closed-loop learning from cognition to action in the virtual world, and then seamlessly migrate to the physical world.

Table of Contents

Terms

1 Overview of Simulation Testing

• 1.1 Simulation Test Phase and Collaboration
  • 1.1.1 Simulation Scenario Modeling Technology
  • 3DGS High-Fidelity Simulation Scenario Reconstruction Technology
  • Simulation Scenario Modeling: 3DGS Case 1
  • Simulation Scenario Modeling: 3DGS Case 2
  • Simulation Scenario Modeling: 3DGS Case 3
  • Simulation Scenario Modeling: 3DGS Case 4 (1)
  • Simulation Scenario Modeling: 3DGS Case 4 (2)
  • Simulation Scenario Modeling: 3DGS Case 4 (3)
  • Simulation Scenario Modeling: 3DGS Case 4 (4)
  • Four Major Development Directions of 3DGS Technology
  • Common Algorithms for 4D Reconstruction and Dynamic Modeling of Autonomous Driving Scenarios
  • Simulation Scenario Modeling: 4DGS Case 1
  • 1.1.2 Physical Sensor Simulation
  • Optical Physical Modeling of Lens Models
  • CMOS Sensor Optoelectronic Simulation
  • LiDAR Modeling: Gaussian Rays and Physical Attenuation
  • Case Analysis: LidarPainter (1)
  • Case Analysis: LidarPainter (2)
  • 1.1.3 Comparison of Vehicle Dynamics Simulation Solutions
  • Case Analysis: New Version of PanoCar 2026 (1)
  • Case Analysis: New Version of PanoCar 2026 (2)
  • Case Analysis: New Version of PanoCar 2026 (3)
  • 1.1.4 X-in-the-Loop Testing and Coordination
  • SIL Case 1:
  • SIL Case 2:
• 1.2 Simulation Scenario Libraries
  • 1.2.1 Scenario Abstraction Hierarchy
  • 1.2.2 Classification of Simulation Scenario Libraries
  • 1.2.3 Data Sources of Scenario Libraries (1)
  • 1.2.3 Data Sources of Scenario Libraries (2)
  • 1.2.3 Data Sources of Scenario Libraries (3)
• 1.3 OpenX Series Standards
• Application Cases of OpenX Series Standards in Simulation Platforms
• Traffic Flow Synthesis Data Platform Case:
• Traffic Flow Synthesis Dataset Case:
• Difficulties in Unstructured Road Simulation
• Case of Overcoming Difficulties in Unstructured Road Simulation:
• 1.4 Relationship between Simulation Testing and World Models
• 1.5 A "Simulation Test + World Model"-Driven Test System
• 1.6 Trend 1 in Simulation Testing/World Models:
• 1.7 Trend 2 in Simulation Testing/World Models:
• 1.8 Trend 3 in Simulation Testing/World Models:
• 1.9 Trend 4 in Simulation Testing/World Models:
• 1.10 Interpretation of National Standard GB/T 47025-2026 Intelligent and Connected Vehicle - Simulation Test Methods and Requirements for Automated Driving Function
• 1.10 List of 48 Test Items in the National Simulation Standard
• 1.11 Simulation Testing and World Model Industry Chain

• 2.1 Overview of World Models
• 2.2 Components of World Models
• 2.3 Technical Indicators of World Models
• 2.4 Core Elements of World Models
• 2.5 Key Technologies and Capabilities of World Models
• 2.6 Evolution of World Model Technology Roadmap
• 2.6 Comparison of World Model Technology Roadmaps (1)
• 2.6 Comparison of World Model Technology Roadmaps (2)
• 2.7 Main Architecture Types of World Models
• 2.8 Development History of World Models
• 2.9 Trends in World Models
• 2.10 Mainstream Application Scenarios of World Models
• 2.11 Mainstream World Model Case 1: Genie3
• 2.11 Mainstream World Model Case 2: Cosmos
• 2.11 Mainstream World Model Case 3: Meta V-JEPA2
• 2.11 Mainstream World Model Case 4: Wayve GAIA2
• 2.11 Mainstream World Model Case 5: SenseWorld World Model
• 2.12 Core Differences and Fusion between World Models and VLA
• From VLA to VLA World Model
• VLA and World Model Fuse to Form WorldVLA Model
• Comparison of Action Model, World Model, and Action World Model
• Technology Route Fusion: VLA + World Model + Reinforcement Learning

3 World Model-Based Simulation: Autonomous Driving

• 3.1 Four Major Values of World Models in Autonomous Driving
• 3.2 Cost Reduction and Efficiency Improvement of World Models in Autonomous Driving
• 3.3 Generation Scenarios of World Models in Autonomous Driving
• 3.4 Key Capabilities of World Models in Autonomous Driving
• 3.5 Deployment Challenges of World Models in Autonomous Driving
• 3.6 Future Directions of World Models in Autonomous Driving
• 3.7 Comparison of World Model Technologies from Major Companies (1)
• 3.7 Comparison of World Model Technologies from Major Companies (2)
• 3.7 Comparison of World Model Technologies from Major Companies (3)
• 3.8 Three Mainstream Technology Schools
  • 3.8.1 Case 1: Pony.ai
  • PonyWorld 2.0
  • 3.8.2 Case 2: SenseAuto
  • SenseWorld
  • 3.8.3 Case 3: QCraft
  • "VLA+ World Model" Unified Architecture
  • 3.8.4 Case 4: Momenta
  • R7 Reinforcement Learning World Model (1)
  • R7 Reinforcement Learning World Model (2)
  • R7 Reinforcement Learning World Model (3)
  • 3.8.5 Case 5: Zhuoyu
  • Multimodal End-to-End VLA World Model (1)
  • Multimodal End-to-End VLA World Model (2)
  • Multimodal End-to-End VLA World Model (5)
  • Ecosystem Partners
  • Vehicle Models Delivered by Zhuoyu
  • 3.8.6 Case 6: Huawei WEWA Architecture (1)
  • 3.8.6 Case 6: Huawei WEWA Architecture (2)
  • 3.8.6 Case 6: Huawei WEWA Architecture (3)
  • 3.8.7 Case 7: NIO
  • NWM
  • 3.8.8 Case 8: XPeng
  • VLA 2.0
  • X-World (1)
  • X-World (2)
  • X-World (3)
  • 3.8.9 Case 9: Li Auto
  • MindVLA-o1 (1)
  • MindVLA-o1 (2)
  • MindVLA-o1 (3)
  • MindVLA-o1 (4)
  • Unified Action Generation: Parallel Decoding with VLA-MoE
  • World Simulator for Reinforcement Learning
  • Physical AI Framework
  • 3.8.10 Case 10: Leapmotor World Model
  • 3.8.11 Case 11: Geely
  • Full-Domain AI 2.0 Based on WAM
  • Key Features of WAM
  • Vehicle Models with WAM
  • G-ASD

4 World Models and Simulation Environment: EAI

• 4.1 Overview and Technical Architecture of EAI
• 4.2 Status Quo of EAI
• 4.3 Core Problem of EAI: Scarcity of Physical Interaction Data
• 4.4 Key Capabilities of World Models in EAI
• 4.5 Three Major Data Challenges in the EAI Field
• 4.6 EAI Data Breakthrough Case 1: Synapath AI
• 4.7 EAI Data Simulation Case 2: Lightwheel
• 4.8 Thought on the Next-Generation Embodied Foundation Model Technology Roadmap
• 4.8 Future Directions of World Models in EAI
• 4.9 World Model Evaluation Benchmark - WorldArena
• 4.10 Overview of Simulation Environments in EAI
• Basic Components of Simulation Environments
• Mainstream Simulation Environments (1)
• Mainstream Simulation Environments (2) - Isaac Sim
• Mainstream Simulation Environments (3) - MUJOCO & Pybullet
• 4.11 Embodied Benchmark Testing
• Common Embodied Benchmark Task Types
• Benchmark Test: CALVIN
• Benchmark Test: LIBERO
• Emerging Benchmark Project: ROBOVERSE
• Benchmark Evaluation Development Trends
• 4.12 Humanoid Robots and EAI Standard System (2026)
• 4.12 Interpretation of Humanoid Robots and EAI Standard System (2026)
• 4.13 Application Case 1 of World Models in EAI: ACE Robotics
• Kairos 3.0 (1)
• Kairos 3.0 (2)
• Kairos 3.0 (3)
• 4.14 Application Case 2 of World Models in EAI: HappyOyster
• 4.15 Application Case 3 of World Models in EAI: LingBot-World
• 4.16 Application Case 4 of World Models in EAI: Genie 3
• 4.17 Application Case 5 of World Models in EAI: HY-World 2.0
• 4.18 Application Case 6 of World Models in EAI: Magic-Mix

5 Mainstream Chinese Simulation Platforms and World Model Solution Providers

• 5.1 GigaAI
• Summary of Core Products
• General Purpose Robot: Maker H01
• Product/Solution Application Cases
• The World's First Comprehensive Data System Driven by a "World Model"
• Embodied Foundation Model: GigaBrain-0
• Embodied Foundation Model: GigaBrain-0.5M* (1)
• Embodied Foundation Model: GigaBrain-0.5M* (2)
• Embodied Foundation Model: GigaBrain-0.5M* (3)
• End-to-End Embodied Foundation Model: GigaBrain-0.1
• Embodied World Model: GigaWorld-0 (1)
• Embodied World Model: GigaWorld-0 (2)
• Embodied World Model: GigaWorld-0 (3)
• Embodied World Model: GigaWorld-0 (4)
• Embodied World Model: GigaWorld-1
• World-Action Model (WAM): GigaWorld-Policy (1)
• World-Action Model (WAM): GigaWorld-Policy (2)
• World-Action Model (WAM): GigaWorld-Policy (3)
• 5.2 GensPi Technology
• Funding and Latest Collaboration
• World-Action Model (WAM): MotuBrain (1)
• World-Action Model (WAM): MotuBrain (4)
• Unified Universal World Model: Motus (1)
• Unified Universal World Model: Motus (4)
• 5.3 AGIBOT
• Open Source Simulation Platform: Genie Sim 3.0
• Unified World Model Platform: Genie Envisioner
• 5.4 51Sim
• SimOne
• SimOne 3.8
• SimOne 4.0 (1)
• SimOne 4.0 (2)
• SimOne 4.0 (3)
• Customer Base of SimOne
• Driver-in-the-Loop (DIL) Solution (1)
• Driver-in-the-Loop (DIL) Solution (2)
• Driver-in-the-Loop (DIL) Solution (3)
• OpenX-based End-to-End Closed-Loop Simulation Platform
• Self-developed 3DGS Hybrid Simulation Engine
• "End-to-End Data-Driven Closed-Loop" Solution
• 5.5 IAE
• Profile
• DeepOcean.AI V3.0
• Simulation Scenario Data Products and Services
• Air-Ground Integrated Simulation Platform
• X-In-Loop(R) Full-Stack Simulation and Testing Toolchain
• X-In-Loop Simulation Solution: Autonomous Driving Digital Twin Simulation Platform
• X-In-Loop Simulation Solution: ECU-Level HIL Testing
• X-In-Loop Simulation Solution: Vehicle-Level HIL (VaHIL(R))
• Model and Software Simulation Testing
• Hardware-in-the-Loop Simulation Testing (1)
• Hardware-in-the-Loop Simulation Testing (2)
• Hardware-in-the-Loop Simulation Testing (3)
• Hardware-in-the-Loop Simulation Testing (4)
• Vehicle-in-the-Loop Simulation Testing (1)
• Vehicle-in-the-Loop Simulation Testing (2)
• Driving Simulator (1)
• Driving Simulator (2)
• Cloud-Accelerated Massive Simulation
• Scenario Workshop
• Software Testing
• Scenario-based Automated Software Testing (1)
• Scenario-based Automated Software Testing (2): Automated Code Verification Tools
• Scenario-based Automated Software Testing (3): Automated Model Verification Tools
• Scenario-based Automated Software Testing (4): Automated System Verification Tools
• Scenario-based Automated Software Testing (5): Software Verification Services
• Scenario-based Automated Software Testing (6)
• Intelligent Connected Vehicle Meteorological Innovation Research Center
• 5.6 Zhejiang PanoSim
• Profile
• Key Technologies
• Application solutions
• Ecological Partners
• Integrated Simulation Test Platform for Autonomous Driving: PanoSim
• Integrated Simulation Test Platform for Autonomous Driving: PanoSim
• High-Precision Vehicle Dynamics Simulation Software: PanoCar
• Latest Version of PanoCar (1)
• Latest Version of PanoCar (2)
• Latest Version of PanoCar (6)
• Autonomous Vehicle Driving Simulator: PanoDrive
• Digital Twin Simulation Platform: PanoTwin
• Digital Twin Autonomous Driving Simulation Platform: PanoTwin-E
• Simulation Test Bench: PanoPilot
• Real-Time Hardware-in-the-Loop Simulation Test Platform: PanoHIL
• 5.7 SaimoAI
• Core Products
• Revenue from Core Products
• Safety Pro
• Cooperation Case 1: Saishi Technology Empowes "Three Consistencies"
• Cooperation Case 2: High-Confidence HIL Test Solution (1)
• Cooperation Case 2: High-Confidence HIL Test Solution (2)
• Cooperation Case 2: High-Confidence HIL Test Solution (3)
• Dual-site Collaboration for Full-chain Verification
• Intelligent Connected Vehicle Closed Test Fields
• 5.8 SYNKROTRON
• Core Technical Advantages
• Technology Empowerment Case 1
• Technology Empowerment Case 2
• Dual Engine: Continuous Learning Framework + World Model for Synthetic Data
• Integrated Autonomous Driving Simulation Platform: OASIS SIM (1)
• Integrated Autonomous Driving Simulation Platform: OASIS SIM (2)
• Integrated Autonomous Driving Simulation Platform: OASIS SIM (3)
• Integrated Autonomous Driving Simulation Platform: OASIS SIM (4)
• Data Acquisition System: OASIS ROVER
• Data Platform: OASIS DATA
• Oasis Connect
• OASIS Traffic
• Oasis BotStream Platform
• EAI Operating System: Dora (1)
• EAI Operating System: Dora (5)
• 5.9 PilotD Technology
• Profile
• Self-evolving Dual-Turbine Driven Data Training Platform
• Highly Physical Synthetic Data Empowerment Cases
• Autonomous Driving Simulation Test Software: GaiA7
• PD Robot Virtual Development and Verification Solution
• PDUranus Intelligent Unmanned Low-Altitude Vehicle Virtual Optimization Platform
• On Cloud Precision Simulation Cloud Computing Platform (1)
• On Cloud Precision Simulation Cloud Computing Platform (2)
• On Cloud Precision Simulation Cloud Computing Platform (3)
• PDRHea(R) Recursive Layered Scenario Generator
• PlenRay(R) Physical Ray Technology
• DiL Solutions
• Medusa High-Confidence Simulation Platform
• Hercules Cost-effective Hardware-in-the-Loop Simulation Platform
• Model Library Refined Classification Platform
• 5.10 Keymotek
• End-to-End Autonomous Driving Simulation Software: aiSim
• aiSim: Overcoming Difficulties in Unstructured Road Simulation
• SiL Verification Solution for the Full R&D Cycle of Autonomous Driving
• aiSim 6 (1)
• aiSim 6 (6)
• aiSim 6 (7)
• Dual-modal Simulation and Test Solution (1)
• Dual-modal Simulation and Test Solution (5)
• High-Fidelity End-to-End HIL Simulation Solution (1)
• High-Fidelity End-to-End HIL Simulation Solution (2)
• High-Fidelity End-to-End HIL Simulation Solution (3)
• High-Fidelity End-to-End HIL Simulation Solution (4): Application Cases
• Keymotek & Jingwei Hirain Autonomous Driving HIL Simulation Test System
• HongKe Builds an EAI Full-Stack Test Solution Covering "Perception-Thinking-Action"
• ROSI Platform (Robotec.ai)
• CoppeliaSim Robot Simulation Platform (Coppelia Robotics) (1)
• CoppeliaSim Robot Simulation Platform (Coppelia Robotics) (2)
• CoppeliaSim Robot Simulation Platform (Coppelia Robotics) (3)
• CoppeliaSim Robot Simulation Platform (Coppelia Robotics) (4)
• CoppeliaSim Robot Simulation Platform (Coppelia Robotics) (5)
• 5.11 Tsing Standard
• Customers of Test Services
• Autonomous Driving Testing: ADAS HIL
• CAN/LIN Bus Test Solutions
• Automotive Ethernet Test Solutions
• CAN Sprite and Ethernet Sprite (1)
• CAN Sprite and Ethernet Sprite (2)
• Customers of CAN Sprite and Ethernet Sprite
• OTA Vehicle-Cloud Integrated Test Solutions
• Steering HIL
• Suspension HIL
• Brake HIL
• BMS HIL
• Full-Dimensional Energy Storage BMS HIL Testing System
• PACK HIL
• Battery Production Line Turnkey Project Cases:
• VCU HIL
• MCU HIL
• BCM HIL
• ZCU HIL
• Comprehensive Simulation Test Solutions for "Electric Drive, Battery, Electric Control" Systems
• Safety Detection Capabilities for "Electric Drive, Battery, Electric Control" Systems
• 5.12 Jingwei Hirain
• INTEWORK VBA
• INTEWORK VBA V3.1.0R
• Newly Launched Software-in-the-Loop (SIL) Testing Platform: INTEWORK-TVM (1)
• Newly Launched Software-in-the-Loop (SIL) Testing Platform: INTEWORK-TVM (2)
• Newly Launched Software-in-the-Loop (SIL) Testing Platform: INTEWORK-TVM (3)
• Newly Launched Software-in-the-Loop (SIL) Testing Platform: INTEWORK-TVM (4)
• Autonomous Driving "Vehicle-Cloud" Evaluation Data Platform
• TestBase VCI Product Series
• TestBase VCI Series Product Upgrade
• DeskHIL Simulation Test Platform (1): Product Features
• DeskHIL Simulation Test Platform (2): Platform Composition
• DeskHIL Simulation Test Platform (3): Software Compatibility
• Next-generation HIL Test Platform (1)
• Next-generation HIL Test Platform (2): Platform Composition
• Next-generation HIL Test Platform (3): Real-time Simulator
• Autonomous Driving HIL Simulation Test Solutions
• Autonomous Driving VIL SYNO Solutions
• Intelligent Cockpit HIL Simulation Test Solution
• Chassis Electronic Control System HIL Simulation Test Solutions
• Powertrain System HIL Simulation Test Solutions
• TPA Smart Lab
• 5.13 PoleLink Information
• Profile
• Intelligent Chassis HiL Test Solution (1)
• Intelligent Chassis HiL Test Solution (2)
• Intelligent Chassis HiL Test Solution (6)
• Intelligent Chassis HiL Test Solution (7)
• Intelligent Chassis HiL Test Solution (8): Steering Motor Simulation Testing
• Intelligent Chassis HiL Test Solution (9): Assembly-Level Steering Testing
• Intelligent Chassis HiL Test Solution (10): Assembly-Level EHB Testing
• Intelligent Chassis HIL Test Solution (11): Suspension Testing
• 5.14 TOSUN
• Simulation Test Solution (1)
• Simulation Test Solution (2)
• Simulation Test Solution (3)
• Chassis HIL Simulation Test Solution
• EMB Automated Test Solution

6 Mainstream Foreign Simulation Platforms and World Model Solution Providers

• 6.1 NVIDIA
• NVIDIA Cosmos 3 (1)
• NVIDIA Cosmos 3 (2)
• NVIDIA Cosmos 3 (3)
• NVIDIA Cosmos 3 (4)
• NVIDIA Cosmos Transfer
• NVIDIA Omniverse (1)
• NVIDIA Omniverse (2)
• NVIDIA Omniverse (3)
• NVIDIA AlpaSim
• NVIDIA Alpamayo 1.5 (1)
• NVIDIA Alpamayo 1.5 (2)
• NVIDIA Alpamayo 1.5 (3)
• NVIDIA Alpamayo 1.5 (4)
• NVIDIA Alpamayo 1.5 (5)
• NVIDIA Isaac GR00T N1.7
• Isaac Lab: GPU-Accelerated Simulation Framework for Multimodal Robot Learning
• NVIDIA Physical AI Data Factory Blueprint
• NVIDIA DRIVE Hyperion 10 Sensor Kit L2++ to L4
• NVIDIA DRIVE L2++
• NVIDIA DRIVE L4 Roadmap
• NVIDIA Halos OS for L4
• AV Business History
• Summary of Enterprises Empowered by NVIDIA
• Automotive Business Ecosystem
• 6.2 UE (Unreal Engine)
• Unreal Engine 5.7
• Summary of Vehicle Model HMI Innovation Cases Empowered by Unreal Engine
• Summary of Digital Cockpit Cases Empowered by Unreal Engine 5 (1)
• Summary of Digital Cockpit Cases Empowered by Unreal Engine 5 (2)
• 6.3 Unity
• Technical Advantages
• Unity China's AI OS 3D Spatial Intelligent Cockpit (1)
• Unity China's AI OS 3D Spatial Intelligent Cockpit (2)
• Unity China's Intelligent Cockpit Solution Customer Base
• Customer Case 1:
• Customer Case 2:
• Customer Case 3:
• 6.4 Wayve
• GAIA-2 (1)
• GAIA-2 (2)
• GAIA-2 (3)
• LA-Pose
• Zero-Shot Practice
• Commercialization Progress and Implementation Plan
• 6.5 Waymo
• Waymo Released World Model Based on Google Genie
• 6.6 Foretellix
• Core Technologies of Data-Driven Autonomous Driving Development Toolchain
• Foretify Data-Driven Development Toolchain
• Foretify Evaluate - Data-Driven Evaluation Framework
• Full-Link Evaluation Process - Foretify Evaluate (1)
• Full-Link Evaluation Process - Foretify Evaluate (2)
• Full-Link Evaluation Process - Foretify Evaluate (3)
• Scalable Neural Reconstruction for Autonomous Driving Development (1)
• Scalable Neural Reconstruction for Autonomous Driving Development (4)
• Cooperation Case 1
• Cooperation Case 2
• 6.7 Hexagon
• Autonomous Driving and Intelligent Driving Simulation Test Platform: VTD (1)
• Autonomous Driving and Intelligent Driving Simulation Test Platform: VTD (5)
• Application of VTD
• Automotive Industry Solutions (1)
• Automotive Industry Solutions (2)
• Automotive Industry Solutions (3)
• Automotive Industry Solutions (4)
• Electric Drive System Engineering Simulation Solutions (1)
• Electric Drive System Engineering Simulation Solutions (2)
• Electric Drive System Engineering Simulation Solutions (3)
• Electric Drive System Engineering Simulation Solutions (4)
• Battery Simulation Solutions
• Body Simulation Solutions
• Component Inspection Solutions
• 6.8 IPG Automotive
• Profile
• Shift-left Test Solution
• Partners
• Cooperation Case 1:
• Cooperation Case 2:
• Product Series
• Product Series: Test System
• Product Series: Software
• Product Series: Hardware
• New Xpack4 Hybrid Solution
• Enhanced Functional Safety and Reliability Capabilities of CarMaker Series
• CarMaker 15.0 (1)
• CarMaker 15.0 (2)
• CarMaker 15.0 (3)
• Integrated HIL Test Solution
• Compact ESC HIL Solution
• SBW-in-the-Loop Test System
• 6.9 dSPACE
• High-Precision Simulation Platform for 3D Scenarios and Physical Sensors: AURELION
• Cloud Simulation Solution: SIMPHERA
• SIL Solutions (1)
• SIL Solutions (2)
• SIL Case 1:
• SIL Case 2:
• SIL Case 3:
• Power HIL
• HIL Cluster Management Solutions
• HIL Case 1:
• Intelligent Chassis Test Solutions (SIL/HIL/Bench)
• VIL: APA Ultrasonic OTA Stimulation and Closed-Loop Test Solutions (1)
• VIL: APA Ultrasonic OTA Stimulation and Closed-Loop Test Solutions (2)
• VIL: APA Ultrasonic OTA Stimulation and Closed-Loop Test Solutions (3)
• New Radar Solution: DARTS ARROW
• RCP Family: High-Performance Integrated Rapid Control Prototype - MicroAutoBox
• SCALEXIO AutoBox: Durable and Reliable Modular Automotive Real-Time System
• BMS 12U HIL, SCALEXIO Rack, SCALEXIO Customized
• New SCALEXIO Essential System
• 6.10 Mechanical Simulation (Affiliated to Applied Intuition)
• Profile and Customers
• SuspensionSim Simulation Software
• TruckSim Simulation Software
• BikeSim Simulation Software
• CarSim Simulation Software (1)
• CarSim Simulation Software (2)
• CarSim Simulation Software (3)
• CarSim Simulation Software (4)
• Basic Usage of CarSim (1)
• Basic Usage of CarSim (5)
• Advanced Features of CarSim (1)
• Advanced Features of CarSim (2)
• Advanced Features of CarSim (6)
• 6.11 Spirent (Affiliated to Keysight)
• Automotive Ethernet Conformance Testing (1)
• Automotive Ethernet Conformance Testing (4)
• 6.12 R&S
• Test Solutions (1)
• Test Solutions (2)
• Test Solutions (7)
• 6.13 VI-grade
• Profile
• Summary of Customer Cases (1)
• Summary of Customer Cases (2)
• Summary of Customer Cases (6)
• Summary of Customer Cases (7)
• Updated Versions of Software Products in Jan 2026
• VI-CarRealTime (1)
• VI-CarRealTime (2): Advantages
• VI-CarRealTime (3): Customers
• Core Advantages of VI-NVHSim
• Configuration of VI-NVHSim Simulator
• Applications and Customer Cases of VI-NVHSim
• Full-Spectrum Dynamic Driving Simulators (1)
• Full-Spectrum Dynamic Driving Simulators (2)
• Full-Spectrum Dynamic Driving Simulators (3)
• HEXAREV Driving Simulator
• Structure of Latest Driving Simulator (HexaRev)
• New HexaRev Motion Platform + HyperDock Cockpit Technology
• Cases of DiM500:
• Dynamic Driving Simulators: DiM400
• Cases of DiM400:
• Dynamic Driving Simulators: DiM150 & DiM250
• Dynamic Driving Simulators: DiM50
• Product Series - Static NVH Driving Simulators
• Product Series - Static Driving Simulators
• Product Series - Compact Full-Spectrum Simulators
• Product Series - Compact HMI Driving Simulators
• Product Series - Compact NVH Driving Simulators
• Product Series - Compact Driving Simulators
• Product Series - Desktop NVH Driving Simulators
• Product Series - Desktop Driving Simulators