|End-to-end Training of Navigation Policies
|LSCM R&D Centre (LSCM)
|Autonomous robots behave and perform tasks with a high degree of autonomy. They rely heavily on technologies in artificial intelligence and robotics. Moreover, depending on their working mode and assigned tasks, robots are equipped with different features to serve their functionalities. Most autonomous robots in the market are designed to work alone and perform simple tasks such as wall-following or random walk-angle changing at home or in office area. To tackle the challenges of coordinating with a team of robots and performing point-to-point navigation in dynamic and complex environments such as airport and warehouse, a more advanced navigation solution using mid-range sensor fusion technologies will be required.
This project aims to develop computational methods for mid-range distance sensing technologies for autonomous robots including
a) End-to-end training of mid-range navigation policies, and
b) Content-based depth estimation for collision avoidance.
This project has been defined by a present need of local industry. With several new technologies incorporated into logistics and supply chain management, the overall productivity will be greatly increased.
|Dr Ka Lun Fan
|Approved Funding Amount:
|16 Mar 2020 - 15 Mar 2021