Multi Sensor Data Fusion, Methods and Problems

Sensors, which monitor the surrounding environment in order to enhance our decisions, play a major role in our lives and contribute to our actions. A single sensor, however, is not capable of providing enough information; therefore, multiple sensors have to be integrated in a way to perform the additional task of interpretation, which may be more helpful and informative than what can be observed using a single sensor. Since the nature of sensor’s functional characteristics can lead to output that contains erroneous measurement readings due to noise, measurement errors, and delays, multiple sensors are needed to confirm the certainty of desired actions. For sensors to work properly, a computational system is required in order to fuse sensor data in a process called multi-sensor-data fusion. This paper presents an overview of multi-sensor-data fusion, using two techniques (Bayes and Dempster-Shafer), with highlights of the techniques’ shortcomings.