A Multidimensional RSSI Based Framework for Autonomous Relay Robots in Harsh Environments

Robotic tele-operation is essential for many dangerous applications, like inspection and manipulation in human hazardous environments. Also, the current state of the art in robotic tele-operation shows the necessity to increase distance between the operator and the robot, while maintaining safety of the operation. Nowadays, delicate manipulation in hazardous environments are mostly performed by robots that were designed for applications such as demining or military purposes, which provide the required level of safety, presenting, however, a series of technology issues, in terms for example of robot localization, cooperation, localization or multimodal human-robot interfaces. In fact, these commercial teleoperated robots normally present the necessity to establish a point-to-point communication between the robot and the base station, reducing its controlling area. This limitation is a difficulty, specially to perform interventions in tunnel environments, such as the one presented at CERN. In this paper a framework for the design of an autonomous relay robot is presented, which allows to have a series of moving stations, in order to extend the communication range between the robot and the operator. The robots are able to navigate safely and to move according to the measured signal strength, in order to maximize the signal throughput between the operator and the robot. The framework is based on different dynamic filtering techniques including Kalman based ones. This allows to predict the signal strength while moving and to react safely to unpredictable environmental changes that might highly affect the signal coverage. The proposed framework has been firstly validated and then successfully deployed on different robotic platforms, while theoretically demonstrated in simulation. Preliminary test results, which have been implemented using the Wi-Fi communication layer, have been tested in the CERN facilities.