Predictive control for vehicular sensor networks based on round-trip time-delay prediction

With the rapid development of vehicular sensor networks (VSNs) technology, the potential use of networked real-time control and automation is enormous and appealing. However, closed-loop control VSNs via the Internet are very difficult to implement practically because of their stochastic nature. One of the biggest challenges in VSNs is how to deal with sensor networks time delay and data loss. The authors investigate the potential of using the predictive control scheme for VSNs based on round-trip time (RTT) delay prediction to overcome the VSNs transmission delay and data loss. The predicted RTT delay is taken as the sampling interval reference for variable-period sampling approach. Modelling the time delay of networked control systems as a non-linear time series, the least mean square (LMS) filter algorithm is adopted to predict online the time delay induced in VSNs. The simulation results show that the LMS algorithm can achieve ideal efficacy if time delay has no variety abnormally. The predictive compensation strategy is proposed to reduce the detrimental effect of stochastic time delays induced by communication networks on control performance. The results of offline simulations via the Internet illustrate that the predictive control scheme based on RTT delay prediction has the potential to overcome the VSNs time delay and data loss.