Towards the development of robot immune system: A combined approach involving innate immune cells and T-lymphocytes

摘要

Mobile robots in uncertain and unstructuredenvironments frequently encounter faults. Therefore, an effective fault detection and recovery mechanism is required. One can possibly investigate natural systems to seek inspiration to develop systems that can handle such faults. Authors, in this pursuit, have explored the possibility of designing an artificial immune system, called Robot Immune System (RIS), to maintain a robot's internal health equilibrium. This contrasts with existing approaches in which specific robotic tasks are performed instead of developing a self-healing robot. In this respect, a fault detection and recovery methodology based on innate and adaptive immune functions has been successfully designed and developed. The immuno-inspired methodology is applied to a simulated robot using Robot Operating System and Virtual Robot Experimentation Platform. Through extensive simulations in increasingly difficult scenarios, the RIS has proven successful in autonomously detecting the abnormal behaviors, performing the recovery actions, and maintaining the homeostasis in the robot. In addition to being multi-tiered, the developed RIS is also a non-deterministic and population-based system.