An Experience with Multi-sensor Robot for Adaptive Human Sensing

Human sensing is vital to multiple applications such as smart homes, assisted living, traffic control, serving robot, rescue operations, etc. For healthy growing, robots are designed for human beings for various purposes such as assisted living, to have a conversation with individuals to avoid depression, a service robot, etc. Since the robot(s) have to work in the close vicinity of humans, it is necessary for the robot(s) to sense the human presence adaptively with the change in environment and also localize itself in an indoor scenario where GPS fails to work. In this paper, we propose a solution based on Analog Ultrasonic Sensor (AUS), Passive Infrared (PIR) sensor, and RGB camera where a robot (on which sensors are mounted) can adapt with changing the environment to sense human presence. Also, the robot can make a decision which sensor to be ruled out from decision making for a particular environment and can localize itself with the help of sonar data in an indoor scenario. Experiments are carried out to apprehend the proposed method in different environments, and results show the efficacy of the proposed method. We have also emphasized on the lessons learned and difficulties faced during the implementation of the adaptive human sensing and localization.