中国具身智能产业发展白皮书（2025版）

White Paper for the Industry Development ofEmbodied Intelligence in China 2025 20264 . . . . . .1 1.1AI+. . . . . . . . . . . . . . .1 1.1.1. . . . . . . . . . . . . . . .11.1.2. . . . . . . . . . . . . . . . .51.1.3. . . . . . . . . . . . .81.2. . . . . . . . . . . . . .121.2.1. . . . . . . . . . . . .121.2.2. . . . . . . . .151.2.3. . . . . . . . . .201.2.4. . . . . . . . . . .231.3. . . . . . . . . . . . . . .261.3.1. . . . . . . . . . .261.3.2. . . . . . .291.3.3. . . . . . . . . .30 . . . . . . . . . . .34 2.1. . . . . . . . . . . . . . . . . . . .342.1.1Internet AIEmbodied AI. . . . . . . . . .342.1.2. . .352.1.3. . . . . . . . . . . . . . . . .362.2. . . . . . . . . . . . . . . . . . . .372.2.1. . . . . . . . . . . . . . . . . . .372.2.2(VLA)Sim2Real. . . . . . . . . . .402.2.3. . . . . . . . . . . . . . .45 . . . . . . . . . . . . . .50 3.1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .50 3.1.1. . . . . . . . . . . . . . . . . . . . . . . . . . . .503.1.2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .513.1.3. . . . . . . . . . . . . . . . . . . . . . . .52 6 3.2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .53 3.2.1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .543.2.2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .573.2.3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58 . . . . . . . . . . . . . . . . . .58 3.3 3.3.1. . . . . . . . . . .593.3.2. . . . . . . . . . . .613.3.3. . . . . . . . . . . .63 3.4Benchmarking. . . . . . . . . . . . .63 3.4.1. . . . . . . . . . . . . . . .653.4.2. . . . . . . . . . . . . . . .66 2030. . . . . . . . . . . . . . . . . . . . . .69 4.1. . . . . . . . . . . . . . . . . . .694.1.1. . . . . . . .694.1.2. . . . . . . . . .714.2. . . . . . . . . . . . . . . . . . .724.2.1. . . . . . . . . . .734.2.2. . . . . . . . . . .734.2.3. . . . . . . . . . .744.2.4. . . . . . . .754.3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .754.3.1. . . . . . . . . . . .764.3.2. . . . . .764.3.3. . . .774.3.4. . . . . .77 1.1AI+ AI 1.1.1 1 20239 20241 20242025+ 1.1 2 20241 2024220246021.1%1.215592030 20253+ 3 420%5 202520252027 1.120232025 +6 + 1.1.2 1 20247 2 20241.3 2024 3 ISO 13849-1 1.020248 2025 1.1.3 1 2025AI 20252027135.02025AI FigureAIHelixOptimus + 2 2025202550%2030400020352026150202511Walker8 2027 +Ultra2025202512App Store+ 412076661.2 3 2024 --2024 1.4 2024Walker S20249Walker S 20248+ 1.2 1.2.1 1 2 20232025 3 -- 1.2.2 20256202720500 20256202320252023915 1 2023915 20256 2023915 202420272024352027502 2 3 1.2.3 1 2022-2025MEMS 2 2025-2027 2025-2027 3 2025-2027 2025-2027Walker 4 +2025-2027240% 2022-2025MEMSAI 2025-2027 5 1000011N 6 2025-2027102050 1000 +AIPOC 1.2.4 1 AGI+ +202320251+N+X1NX50++10008 202492024-2026225225202650100 2 + + 3 2024202620288002002030 Sim2Real200 1.3 -- 1.3.1 C9-- 1C9 1.5 •2025 + •BIGAI •--2000 •20252.0 2 3 2025 20241091.5 1.3.2 1 •BIGAIAGIBIGAIAGIBIGAI-Embodied10 •SAILOpenXLab Embodied--VLA •-- •Panoramic Sim85% -1PB 2 3 1. 2.S-hand95% 3.Learning fromDemonstrationAI Lab +RSSICRA IROS 1.3.3 1 2025 •35202510+ •2026 •5+ • 2 1.6 1. 2. 3 AIGC 1.6 1.3 2.1 2.1.1Internet AIEmbodied AI Internet AI Internet AIInternet AI Internet AI Embodied AI 2.1 • • • 2.1.2 Internet AIAI Embodied AI Internet AI Embodied AI StGΩAtSt+1 Internet AIEmbod-ied AI2.2 AIAI 2.1.3 Sim2Real2.3Sim2Real ΘSASΘAI 2.2 2.2.1 1 2 RGBRGB/IMU/ 3 // // 4 5 IMUIMUIMU 2.2.2(VLA)Sim2Real 1 2 DeepSeek-R1OpenAI o1RLRL MDP(S, A, P, R, γ)(S)(A)(P)(R) (γ)π2.6 Vπ(s)sQπ(s, a)saQ Q-LearningPPO 3 AIGCMOOParetoOptimality fi(x)iwiw Model Predictive Control, MPC MPCtNutt+ 12.7 4 Multimodal DataFusion2.8 5 2.9 •(Initialization) • •(The Evolutionary Loop) –(Selection) –(Crossover) –(Mutation) • 2.10PSO- •t(xt) • –(pBest)pBest–(gBest)gBest •(vt+1)xt(xt+1) PSO 2.2.3 1 -2.11-- 2.11 2 2.12 2.12 3 2.13 2.13 B 4 2.14PID 2.2 2.14 PIDPID - / 3.1 3.1.1 / :2.02.0 202592.02.03.12.0812 3.1.2 Walker S1 20255GSPS3.2 Walker S1BrainNetDeepSeek-R1 Walker S1- // 2025126UMIROU93D 1438TPM3.380%10WRCA 3.2 ++ 3.2.1 2025++ 1 20259JDDiscovery-20252.0++3.4---242.0 2 3.5 2025111003 3:++ 202510 203.63620650 / 3.2.2 / 202583.7GalbotGroceryVLAGraspVLA 9/300+ 200050030%-40% 3.2.3 XMAN-F1 20257XMAN-F13.8XMAN-F1ProS 3.3 3.3.1 100020001790%50%3.9 5008194460% 95% 500AI 580%2.572621098% 160++AI1431.83293% 3.3.2 16.22203205325%12024278AI3.11++ X20 SLAM2022 3.3.3 [13]20261AI 342120.650% 3.4Benchmarking benchmark 3.4.1 20247AI20255 MG400MG400190mm×190mm±0.05mm6MG400PCB 430% my-Cobot 280myCobot 280Jetson Nano 3.4.2 [18]5G WAIC 2025GRx[19]30603000300[20] 20 2030 2030 4.1 4.1.1 1. AI5-10 COCO-LocoGibson Env 2. 30%50% 3. Sim2RealNVIDIAIsaac Sim 4.1.2 1. 70%15%