Implementation of a reliability test protocol for a multimeasurement sensor dedicated to industrial applications of the Internet of things

The Internet of things (IoT) is considering incorporating many wireless sensors into the same object to make it smart. To satisfy the main applications of IoT (smart building, smart city and industry 4.0), Multimeasurement Wireless Sensor (MWS) capable of simultaneously measuring several different environmental data (temperature, pressure, vibrations, etc.) are increasingly marketed. In these sensors, the same physical principle is used to acquire various data. From this, we propose in this paper a reliability test protocol for MWS. This study is more important in IoT applications since decision-making based on erroneous data from the sensor can lead to harmful, incorrect and costly actions. Depending on the different quantities measurable by the sensor, it is set up a test platform and the used analysis method is that by comparison of results with specific sensors to the measurement of a single physical data. Among the performance criteria of the sensors, it is considered in this paper those related to the stability of the sensor, its response time, its reaction in hard environments and finally the inter-influence between the different controlled data. To evaluate the correlation between the different physical quantities under control, a method called the dial method, inspired by the Taguchi method is proposed. The method is to define an acceptable tolerance and analyze the behavior of one measure when the other measurand fluctuates between two previously defined extreme values. The MWS considered in this work is the XDK110 sensor designed and marketed by Bosch Rexroth. This sensor allows measuring eight different physical quantities. A comparison between the values obtained from the test platform with those proposed by the manufacturer is made. It is observed, for example, for a normal temperature level, that the measured magnetic field is highly sensitive to strong vibrations whereas in the case of high temperatures the indicated value is quite close to the standard amount. From these observations, solutions are suggested to the designers and users, one of the main being the development of correction algorithms to be integrated into the MWS. (C) 2019 Elsevier Ltd. All rights reserved.