Mobile robot dynamic motion planning to avoid collisions with humans
Insinööritieteiden korkeakoulu | Master's thesis
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.
Master's Programme in Mechanical Engineering (MEC)
AbstractIn various scenarios, e.g. on a construction site, in a shopping mall, or on a street, a mobile robot confronts pedestrians from time to time. Consequently, future paths of the robot should avoid pedestrians to keep them safe. Research on this topic has gradually increased. Many processes of human-robot interactions, e.g. humans and robots walk towards each other, move forward together, or meet on the road, are similar to the scenario above. Therefore it is necessary to generate one safe path before the robot sets up, namely planning one global path from the starting point to the goal site. To achieve a continuous, safe, and efficient path, it is essential first to predict the pedestrians' future path. In the thesis, the method of causal human trajectory prediction is adopted to predict the future path based on the discontinuous previous clues, while CNN as well as RNNs is emphasized to encode the history information to predict future trajectories. For path planning, A star as the algorithm is utilized to search for one path to detour the obstacles.
Thesis advisorZhao, Wenshuai
path planning, trajectory prediction, object tracking, object detection