Internet of Things Based Monitoring and AI Based Testing Approach for Sensor Activity Testing of Embedded Systems

Author Name(s): Swapnili Karmore, Anjali Mahajan, Pravin Karmore
Author Email:


The Internet of Things has been applied in many of the industries. Huge and large industries demand human less monitoring systems. Embedded system testing process proved as most critical and complex process in industries. In Embedded system testing, continues monitoring and testing of input, outputs and sensory activity is an essential requirement, each device is such typically equipped with a number of sensors, hence the multiple electronic control units are connected in order to provide control of the system. Normal Software testing is not suitable for embedded systems. Software as well as Hardware testing is required in case of embedded system, Testing of Embedded systems demonstrated through several simulation experiments. In this article, we propose small case studies for designing and development of testing framework via AI and IOT based approach, through extensive experiments testing of custom build an embedded system exhibits the variety of results proving its designing accuracy. The analysis is performed when data is generated synthetically as well as by actual developed embedded system. It also achieves increased efficiency, demonstrating the classification of valid and invalid data.


Now a day’s Internet of thing is the most popular way of monitoring and managing the work associated in large industries. Embedded systems are known as the most demanding and a popular system, testing of embedded system’s functionality is very complex and critical task. It includes real time environment monitoring in large buildings, medical equipment’s functionality testing and accuracy in systems used in military and industry applications. Among several testing mechanism new approaches used for embedded system testing are Artificial Intelligence based. Intelligent Networks, that is responsible to monitor as well as to note the testing results of the systems. An embedded system requires a number of sensors to indicate its different running states. Misbehavior of input mechanism can lead failure of output mechanisms. The sensors are responsible to show the behavior of the embedded system, whether the function which makes its safety critical are performing well or not. Artificial intelligence approach is used for testing of embedded system based on the classification problem used for nonlinear systems.


The work described in this research is monitored via Internet of Things. The dashboard is developed for Admin side and android based application is developed toward User side. The Embedded system, Oil level detection system testing is performed via ANN. In the version one the overall efficiency obtained from training is 99 %, testing 98.1% and validation is 97.1% in the version two. From the above results the overall efficiency obtained from training is 99%Testing 99% and validation is 97.6%. From the table and the dataset value it is observed. The safety critical function, accuracy its performance and error rate can be successfully calculated via LM algorithm. Classification of valid and invalid classes leads to avoid failure cases which improve efficiency of the overall system. Development of such environment avoids, use of real bulky, costly and complex embedded system for testing. Development of an artificial intelligence approach is competent enough to provide a common platform, where application specific and differential characteristics embedded systems can be tested successfully. Internet of things makes it human less platform independent system

377 total views, 1 views today

Download PDF File

About the author: admin