Improving coverage control laws for multi-robot systems

Consider the following applications:

• Monitoring the temperature/current or oil spills in a part of the ocean.
• Attempting to ascertain the presence of nuclear radiation in a particular region.
• Deploying a network of unmanned (UAVs) in a reconnaissance mission to measure enemy presence in a region.

In each, there is a need to distribute autonomous agents functioning as sensors to optimally cover an unknown environment. Algorithm design for such agents usually faces two challenges. The first is location optimization — to resolve how the agents should be ideally distributed in order to achieve optimum sensing capabilities. The second, and more important, is to ascertain the requirements in communication in order to achieve the desired optimum coverage configuration in a robust manner (for instance intervals between messages, communication topology).

Rihab Abdul Razak, a research scholar with the IITB-Monash Research Academy, is working on a project titled, ‘New models and algorithms for decentralized and adaptive coverage control in multi-robot systems’ to help resolve these challenges.

He aims to study advances in decentralized coverage control. This is an important coordinated control problem wherein a set of robots identify optimal techniques to cover an unknown environment for the purpose of sensing.

Says Rihab, “The phenomenon to be sensed by the robots is described by a mathematical function over the region to be covered. Each robot is capable of measuring the value of the function at its position. Controllers for mobile robots have been developed so that the robots realign themselves in an optimal sensing configuration. The problem is solved by optimizing the locational optimization cost function with respect to the agent positions, and developing controllers for the robots so that they converge to the optimum of the cost function. The controllers are decentralized so that the controllers for each robot in the group work based on information local to the robot and do not require global information.”

Most of the work done in this field so far uses simple agent models to design controllers. “We aim to design controllers for more realistic agent models,” explains Rihab. “Also, we focus on adaptive algorithms where all the information about the environment may not be known, and the algorithms developed need to adapt to these unknown quantities.”

The Academy is a collaboration between India and Australia that endeavours to strengthen scientific relationships between the two countries. Graduate research scholars like Rihab study for a dually-badged PhD from both IIT Bombay and Monash University, spending time at both institutions to enrich their research experience.

Says Prof Murali Sastry, CEO of the IITB-Monash Research Academy, “Rihab’s work has wide-ranging applications, including rescue and recovery, radiation and nuclear spill detection, underwater oil exploration, etc. The goal is to improve existing coverage control laws.”