A soft sensor for energy efficient application of wireless networked control systems

Recent advances in wireless networking technology and proliferation of industrial wireless sensor have led to an increased interest in using wireless networks in NCSs. The main advantages of these Wireless Networked Control Systems (WNCSs) are the reconfigurability, easy commissioning and the installation in places where cabling is impossible. Despite these advantages, there are two main problems which must be considered for application of WNCSs. One problem is the sampling period constraint of the industrial wireless sensors. This problem is related to the energy cost of the wireless transmission, once power supply is limited, which unable the use of these sensors in several closed-loop controls. The other concern in WNCS is the energy efficiency of devices. As the sensors are powered by batteries, it is expected the lowest possible consumption to extend battery lifetime. This paper proposes a model based soft sensor to overcome these problems in WNCSs. The goal of the soft sensor is generating virtual data allowing an actuation on the process faster than the minimum sampling period available for the wireless sensor. Experimental results in a WNCS testbed shown that the soft sensor provided reliable virtual data for the multi-rate controller, overcoming the sampling period constraint and enabling the application of industrial wireless sensors in WNCSs. Additionally, the results shown that the soft-sensor is energy efficient. By increasing the wireless sensor sampling period, while maintaining control performance, the soft sensor can also be applied for saving battery of the industrial wireless sensors.