Preliminary tests conducted at TTCI showed that it is possible to achieve error rates that are close to commercial train WIM stations using the accelerometer based WSN system. The installed cost of the WSN system is about 20-30 times less than the conventional train WIM stations.

 

     Sensor configuration for preliminary testing - Red blocks are the sensor locations.

                                 Acceleration response from 10 bogies

 

Using the preliminary test results, an average mean absolute relative error (MARE) of  3.8% is achieved using the car weight predictions of 113 cars of a loaded train.  It should be noted that only one sensor is used for preliminary testing and curve fitting method is not used to obtain accelerations. Using multiple sensors and a curve fitting method, a 50% reduction in error rates is likely to be observed (Bajwa et al. 2013).