Research Projects

The advancement of technology is focusing on smart cities, where intelligent robotic systems can improve human life. One crucial component of this is autonomous transportation, which is currently suitable for low-traffic environments like highways and wide roads. However, this proposal aims to investigate autonomous navigation in crowded environments, such as accident sites where people gather around an ambulance. The challenge lies in detecting and understanding human intent to avoid collision with the crowd. This project aims to develop and integrate a novel human intent prediction strategy with current deep reinforcement learning strategies for autonomous navigation in crowded traffic routes. The project will be executed in two phases: Phase 1 will develop a situational awareness and assessment module for a crowded environment, creating datasets for human intent prediction and algorithms for assessing the crowded environment. Phase 2 will develop and test the autonomous navigation module on autonomous ground vehicles in a crowded environment, including motion planning and collision avoidance algorithms. The project requires expertise in autonomous vehicles, computer vision, learning, motion prediction, path planning, and collision avoidance. The proposed algorithms can be applied to any crowded environment consisting of a heterogeneous mix of entities such as vehicles, humans, and human-operated mechanized mobility systems. The outcomes of the project could improve the autonomy of existing autonomous vehicles and help integrate autonomous vehicles with manned traffic for future smart cities.