Multi-robot Navigation

Overview

I participated in a military project for developing a team of mobile robots as a researcher in Korea Institute of Science and Technology (KIST). The multi-robot team consists of four six-wheeled skid-steering robots. The team uses a nonlinear model predictive control (NMPC) algorithm to control its formation and navigation in unstructured outdoor environments. I worked with the real robots to test the control algorithm. My main duties were integrating sensors (e.g., DGPS, IMU, laser range sensors, and optical sensors) and communication devices (e.g., wireless LAN, Bluetooth, and Zigbee) and testing them for autonomous navigation. I tuned parameters for control and post-processing of sensor data. Also, I operated a central command unit to control the robots. I worked on resolving practical problems of sensors (e.g., signal interferences among wireless devices, inaccuracy of DGPS and IMU sensors in particular situations, and parameter settings of optical sensors for different environments) and hardware (e.g., robot assembly and maintenance). I wrote a patent document for a localization sensor developed in my team. The sensor integrates an optical sensor and an inertial measurement unit to accurately estimate a robot's position without an aid of GPS.