Design and developement of a sensor-based autonomous navigation system requires that a mobile robot is capable of exploring (interacting with its environment) and planning how to travel through an uncertain environment and executing the plan to accomplish the missions, e.g. mapping a region, gathering data, surveillance (target seeking, following or pursuit, and border patrol) in potentially hostile environment. To meet the requirement, the robot needs to solve the path/motion planning problem, i.e. generate a static or dynamic obstacle-avoiding path/trajectory that is optimal with respect to one or more kinematic or dynamic criteria.
The following are some of our aims:
1. proof-of-concepts experiments for mobile robot navigation on unknown indoor environments
--3D mapping using rotating laser range finder based on DT +linear simplex for map alignment
--Waypoint navigation system for limit cycle avoidance in nonconvex obstacles (e.g. U-shape or deadend) in polar coordinate to complement goal-directed reactive navigation
--surveillance mobile robot performing wall following and human detection
2. Path planning algorithms search for the optimal or near-optimal paths linking the given initial and target configurations that optimize an objective function (eg the path length) and satisfy the hard and soft constraints imposed on the vehicle, such as avoiding obstacles or response to changes in the environment. Now we develop soft computing based approaches for mobile robot path planning for data gethering in wireless sensor networkand navigation.We develop an efficient evolutionary path planner for data-collection problem using data mule in wsn or using uav for mapping a region, which is a variant of Traveling Salesman Problem with Neighborhood (TSPN).