What are we solving?

A mobile robot is placed in an unknown environment and must track a moving target. The mobile robot must simultaneously map the locations of the radio sensors, the mobile target, and localize itself within the environment through received signal strength measurements while also avoiding all obstacles in its path.

Abstract

One of the most fundamental problems in the field of mobile robotics is the ability for a mobile robot to both localize itself and generate a map in an unknown environment using only external sensor measurements. A particular challenge in the simultaneous localization and mapping area is that of a cost effective and modular solution that can be used in an indoor environment. Even more complexity can be added to this challenge if one extends the scope to include solving the mobile target tracking problem as well.

The mobile target racking problem is as such; the mobile target and follower robot are both placed in an unknown environment where neither the pose (position and orientation) of either are known a priori. The follower must then match the unknown trajectory of the mobile target using only external sensory measurements. This work approaches this problem using range only measurements generated by a network of wireless radio sensors placed randomly throughout the robot's operating environment. A custom built radio transceiver is mounted on the follower robot to gather these measurements. The path of the mobile target in the operating environment is completely random and represented by another radio sensor in the wireless sensor network.

A conventional Extended Kalman Filter Simultaneous Localization and Mapping (EKF-SLAM) algorithm will be used to process the wireless measurements taken by the radio receiver. The EKF-SLAM algorithm will estimate the pose of the follower robot, the position of the mobile target, and generate a map of the environment by estimating the position of the wireless sensors placed in the operating environment. A breitenberg object avoidance algorithm will be implemented utilizing the sonar sensors on the mobile robot to navigate around obstacles. A PI (proportional integral) controller will be used to calculate the necessary velocity for the follower robot to track the mobile target within a minimum safe distance. We will first simulate this work using a commercial robotics simulator V-REP (Virtual Robot Experimentation Platform) and then implement it using a mixture of custom and pre-built hardware.

Click here to take a look at the high level system overview.

Nicolas Auth

is a senior electrical engineering student pursuing a Bachelor's degree from Bradley University, Peoria, Illinois, USA. His interests include studying mobile robotics and experimenting with audio systems.

Grant Hovey

is a senior electrical engineering student pursuing a Bachelor's degree from Bradley University, Peoria, Illinois, USA. He is currently employed at Komatsu America Corporation.