Multi-robot motion planning without a central coordinator usually relies on the sharing of planned trajectories among the robots via wireless communication in order to achieve predictive collision avoidance. Path planners found in the literature that feature this scheme usually boast levels of performance comparable with their centralized counterparts. However, in practice, communication tends to be unreliable and lead to significant performance degradation, which may eventually result in collisions among robots.

This thesis proposes a data-driven decentralized approach for multi-robot motion planning in dynamic environments. Instead of requiring robots to share their intentions with each other, a model based on recurrent neural networks (RNNs) is used to predict them. This model is trained on data obtained with a centralized sequential model predictive control (MPC)-based motion planner in which intended trajectories of neighboring robots are available while planning. The learned model can be efficiently run online and provide accurate predictions for each robot in the environment based on a horizon of past observations of all robots' states. This model is then incorporated into the MPC planning framework so that it may be run in a decentralized manner. It is finally shown in simulations that the proposed approach achieves a similar level of performance to its centralized counterpart.

Safe navigation in unknown environments is a challenging task for autonomous Micro Aerial Vehicle (MAV) systems. Previous works generally avoid obstacles by assuming that the environment is static. The purpose of this thesis work is to develop a MAV system that can navigate autonomously and safely in dynamic environments. We present an onboard vision-based approach for the avoidance of moving obstacles in dynamic environments. This approach uses a state-of-art visual odometry algorithm to estimate the pose of MAV and an efficient obstacle sensing method based on stereo image pairs to estimate the center position, velocity, and size of the obstacles. Considering the uncertainties of the estimations, a chance-constrained Model Predictive Controller (MPC) is applied to achieve robust collision avoidance. The method takes into account the MAV’s dynamics, state estimation and the obstacle sensing results ensuring that the collision probability between the MAV and each obstacle is below a specified threshold. The proposed approach is implemented on a designed experimental platform that consists of a quadrotor, a depth camera, and a single-board computer, and is successfully tested in a variety of environments, showing effective online collision avoidance of moving obstacles.