Open Intelligent Electric Vehicles Research Center (OIEV-RC)
Location：
The MUST LIU’s Innovation and Technology Center, ITC-14
Brief introduction：
Faculty of Innovation Engineering, along with its industry partner AutoCore, a pioneer in the area of autonomous driving technologies, has established an Open Intelligent Electric Vehicles Research Center (OIEV-RC). Our research objectives are to focus on developing the next-generation autonomous driving technologies which allow multiple agents to collaborate for driving tasks and for the performance of the models and functions to be incrementally improved over time through STEP perpetual learning and collective intelligence. The ultimate goal is to realize collective intelligence solutions on wheels.
Currently, OIEV-RC owns Macau’s first complete demonstration autonomous vehicle modified from a conventional electric vehicle (Model: VOLKSWAGEN ID.3 PURE A/T VE). For the research and testing of autonomous driving technology, this autonomous vehicle is designed with open accessible software and hardware interfaces that allow the updates and improvements of the system from any third-party plug-ins algorithm. Currently, the vehicle has obtained permission from the Macau Transport Bureau to carry out road testing.
The vehicle is equipped with sensors including: a forward-facing camera, three wide-angle cameras, a forward-facing solid-state lidar, and a top 360° lidar. A total of six sensors are used to perform sensing tasks, and an inertial navigation system (INS) It is used to achieve precise positioning. CPU-based industrial computers are used to execute algorithms such as perception result fusion and decision-making. GPU-based industrial computers are used to execute perception algorithms and planning algorithms.
Currently, the deployed autonomous driving system supports avoidance and detection functions of vehicles and pedestrian. The vehicle detection module fuses the depth detection information of the laser radar (LiDAR) and the semantic recognition information of the camera. By fusing the road topology information and traffic rules, the tracking module can predict the driving route of the vehicle. When there is a risk of collision between an unmanned vehicle and other vehicles, the system will slow down or even make an emergency stop to avoid a collision. Moreover, the system can also accurately identify the real-time location of pedestrians in three-dimensional space and predict the trajectory of pedestrians on the road, so that to slow down or even make an emergency stop to ensure pedestrian safety.
Through such an open and intelligent platform, we will enable the researchers to conduct their work at the bleeding edge, to offer an open canvas for the brilliant minds, and to tackle the toughest challenges the industry is facing, to equip the students with best in class technology and access to the widest spectrum of career opportunities. This is a new beginning for which we are full of hope and expectation - the best is yet to come!
Industry partner:
AutoCore Intelligent Technology (Nanjing) Co., Ltd. (AutoCore) is now a privately held, venture capital-backed automotive company and a founding senior member of the AutoWare Foundation, the world's leading autonomous driving open source software project. The main business is to realize the safety middleware and SoC-independent automotive-grade computing platform required for autonomous driving in a low-cost, low-power, scalable and reliable manner through open source software.
Major research projects:
Zhou, B., Wang, P., Wan, J., Liang, Y., & Wang, F. (2023). A Unified Multimodal De-and Re-Coupling Framework for RGB-D Motion Recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). (CAA-A, CCF-A)
Zhu, Z., Yang, L., Li, N., Jiang, C., & Liang, Y*. (2023). UMIFormer: Mining the Correlations between Similar Tokens for Multi-View 3D Reconstruction. ICCV 2023. (CCF-A)
Yang, L., Zhu, Z., Lin, X., Nong, J., & Liang, Y*. (2023). Long-Range Grouping Transformer for Multi-View 3D Reconstruction. ICCV 2023. (CCF-A)
Liu, B., Pu, Z., Pan, Y., Yi, J., Liang, Y., & Zhang, D. (2023). Lazy Agents: A New Perspective on Solving Sparse Reward Problem in Multi-agent Reinforcement Learning, ICML 2023. (CCF-A)
Lin, X., Gan, J., Jiang, C., Xue, S., & Liang, Y.* (2023). Wi-Fi-Based Indoor Localization and Navigation: A Robot-Aided Hybrid Deep Learning Approach. Sensors, 23(14), 6320.
Yu, X., Liang, Y.*, Lin, X., Wan, J., Wang, T., & Dai, H. N. (2022). Frequency Feature Pyramid Network with Global-Local Consistency Loss for Crowd-and-Vehicle Counting in Congested Scenes. IEEE Transactions on Intelligent Transportation Systems (TITS) (CAA-A, CCF-B)
Li, J., Cheang, C.F., Liu, S., Tang, S. Li, T. & Cheng, Q. (2024). Dynamic-TLD: A Traffic Light Detector based on Dynamic Strategies. IEEE Sensors Journal. DOI: 10.1109/JSEN.2024.3352830
