Formal modelling and verification approach for improving probabilistic behaviour of Robot Swarms

Robot Swarms (RS) is an approach that leads to the combined functioning of multiple robot systems to achieve a desired collective behaviour. Robot Swarms possess multiple benefits e.g. RS employment in calamity salvage missions: Swarms of robots of various sizes could be sent to the spots where salvage specialists can't reach securely to detect the presence of life. Swarms can be utilized as a part of military to frame a self-governing armed force. Apart from a large number of advantages, the biggest challenge for RS is its non-predictable behaviour. Architectures and protocols for working of swarms of robots have already been designed and implemented, but to improve the non-predictability of swarms either experiments with robots are to be carried out or computational simulations are performed. However, with these approaches, a deep analysis of all the possible behaviours is very difficult. To improve the non-predictability of RS we used a potent technique, formal verification to confirm that it is bug free. Researchers have realized that computer simulations alone are not capable of evaluating a system or protocol against all possible scenarios. This has encouraged us to employ model checker tools like UPAAL to evaluate Path Planning Algorithms and Alive Human Navigation through RS in disastrous situation. After verification of probabilistic behaviour of swarms, finally, we can safely claim that the collective working of a robot swarm will be free of any undesirable behaviour.