Energy-efficient sensing in robotic networks

In this paper, we propose a distributed data collection algorithm for robotic networks, which exploits the integration between compressed sensing (CS) and collaboration of mobile robots. Based on the fact that the mobile robots can move into random positions in a sensing area that need to be observed, at a time instant, data collected from a certain number of connected robots can create a sparse random projection so-called CS measurement. At a sampling time, the robots collaborate and share their sensory readings to each other within their transmission range. This linear combination is called a CS measurement to be stored at each distributed robot. The greater the sampling times, the greater the number of CS measurements generated at each mobile robot. Each distributed robot can reconstruct sensory readings from all positions in the area based on the number of CS measurements that is much smaller than the number of positions in the sensing field. We analyze and formulate power consumption for data transmission in such networks. We also analyze the number of mobile robots, the trade-off between the convergence time and robot transmission range and suggest an optimal range for the mobile robots to consume the least power. (C) 2020 Elsevier Ltd. All rights reserved.