Time-critical missions are among the most important classes of tasks yet to fully benefit from agile autonomous micro aerial vehicles. At the same time, such missions require developing algorithms that enable these robots to navigate safely, and make effective decisions on resource-constrained hardware. In this talk I will focus on algorithms for two classes of trajectory optimization problems for quadrotors. First, we shall look into the structure of perception-aware motion planning constraints that allow a vision-driven robot to traverse a given geometric path in minimum time while maintaining accurate state estimates at all times. Second, where only part of the structure can be identified, we shall turn to a fusion of model-based and learning-augmented approaches to synthesize dynamically-feasible trajectories at a fraction of the computational cost of traditional methods.
Igor Spasojevic is an Assistant Professor in the Bourns College of Engineering at the University of California, Riverside. Prior to UCR he was a postdoc in the General Robotics, Automation, Sensing, and Perception (GRASP) laboratory at the University of Pennsylvania. He received his Ph.D. from the Massachusetts Institute of Technology, where he worked on algorithms for perception-aware motion planning for agile micro aerial vehicles. His research spans various types of optimization problems in robotics, recently with a focus on trajectory planning and autonomous exploration.