Reliable Robot-Flock-Based Monitoring System Design via A Mobile Wireless Sensor Network

• Main Authors: Peng Zeng, Jiahong He, Bingtuan Gao
• Format: Article
• Language: English
• Published: IEEE 2021-01-01
• Series: IEEE Access
• Subjects: Power cable tunnel; monitoring system; spherical robot; mobile wireless sensor network; leader-follower strategy; formation optimization
• Online Access: https://ieeexplore.ieee.org/document/9382970/

The reliable distributed monitoring system provides the real-time video observation from the power cable tunnels when the preinstalled communication system is disabled in an emergency. The spherical robotic prototypes are designed as the sensor node based on the analysis of the robot dynamic model. A novel self-constructed mobile wireless sensor network (MWSN) is proposed to realize the communication and positioning functions through the fusion of the Ad-Hoc and ultra-wideband (UWB) approaches. The Ad-Hoc network consists of multiple robot nodes as sensor carriers and relay stations to transmit large amounts of data from the accident scene to the monitoring terminal. The UWB network consists of UWB tags and bases on each node and implements the leader-follower (LF) strategy to create the robot formation. In order to maximize the monitoring area and maintain the reliability of the network, the traditional LF strategy is optimized based on the <italic>N-X</italic> algorithm subjected to the constraints of node number, Ad-Hoc communication and UWB positioning range, and the power tunnel boundaries. A self-constructed MWSN consisting of five robotic nodes (<inline-formula> <tex-math notation="LaTeX">$N=5$ </tex-math></inline-formula>) is capable of covering 40 m of power cable tunnel, when the value of <inline-formula> <tex-math notation="LaTeX">$X$ </tex-math></inline-formula> is selected as 3. The network failure probability reduces from 73.5&#x0025; (<inline-formula> <tex-math notation="LaTeX">$X=0$ </tex-math></inline-formula>) to 0.2&#x0025; (<inline-formula> <tex-math notation="LaTeX">$X=3$ </tex-math></inline-formula>) when the malfunction probability of sensor node is 23.3&#x0025; due to the accident condition in the tunnel. Finally, the simulation and experimental results show that the optimized LF algorithm increases 20.2&#x0025; coverage of the network in the curve formation. The network transmits a monitoring video streaming with a <inline-formula> <tex-math notation="LaTeX">$320\times240$ </tex-math></inline-formula> pixel resolution and a delay less than 150 ms. The real-time video observation from the accident scene is significant to formulate the emergency countermeasure and investigate the direct cause of the accident.