AGV-Assisted Adaptive Cooperative Transmission for State Estimation in Industrial IoT Systems

The wide application of Internet of Things (IoT) technology in industrial automation promotes the emergence of industrial IoT systems, in which state estimation plays a crucial role in conjecturing system states with sensory data delivered over wireless channels. In this paper, we propose an automated guided vehicle (AGV)-assisted adaptive cooperative transmission scheme to minimize the mean square error of state estimation at a low energy cost. Specifically, a novel performance index, estimation gain, is introduced to evaluate the benefit of scheduling one sensor for estimation error reduction. Then, sensor scheduling and data transmission are jointly optimized to minimize the time-accumulated estimation error, which is challenging to directly solve due to the unclear impact of imperfect transmission on estimation performance. To this end, an estimation gain-based algorithm is designed to determine the scheduled sensors. Besides, an iterative algorithm is designed to solve the adaptive cooperative transmission problem. Simulation results show that the proposed scheme outperforms benchmark schemes in reducing estimation error and energy consumption.

I. Introduction

Internet of Things (IoT) technology have been widely used in actual industrial systems such as equipment predictive maintenance, production efficiency control and intelligent logistics, promoting the emergence of industrial IoT systems [1], [1], [2], [3], [4], [5], [6]. In industrial IoT systems, sensors are widely deployed to collect sensory data, which is then used by the remote controller to perform state estimation to conjecture the operating status of industrial process [7], [8], [9]. Therefore, the performance of state estimation heavily relies on the transmission performance, since to information exchange through the network [10], [11], [12]. However, the unreliability of wireless channels and the complex and dynamic industrial transmission environment result in poor transmission performance, making it challenge to achieve high-accuracy state estimation for industrial IoT systems [13].