Methodology to Recognize Vehicle Loading Condition – An Indirect Method Using Telematics and Machine Learning 2019-26-0019
Connected vehicles technology is experiencing a boom across the globe. Vehicle manufacturers started to have telematics devices which leverage mobile connectivity to pool the data. Though the primary purpose of the telematics devices is location tracking, the additional vehicle information gathered through the devices can bring in much more insights about the vehicles. Cloud computing is enabling standard connected vehicle offerings, and on the other hand machine learning and data analytics enables rich customer experience at a very limited cost.
From a fleet owner perspective, the revenue and the maintenance costs are directly related to the loading conditions of the vehicle. This could help in efficient planning on vehicle sharing, drive mode selection and proactive maintenance. A common approach to vehicle load condition detection is by using costly sensors. This paper explores a possibility of detecting vehicle load conditions without making use of any additional sensors. Instead, a supervised machine learning model is developed to recognize real-time loading condition, by analyzing vehicle driving behavior.
This paper is covered in four sections mainly. The first section describes data sources, parameters, sampling and recording of relevant parameters. Parameters considered are from in-vehicle controller area network (vehicle and speed, driver inputs, gear position, and torque) and vehicles response from the inbuilt accelerometer. Statistical and domain-specific feature extraction from time series signal is described in section two. Different machine learning models are compared in section three and the final implementation details covered in section four. Primary vehicles under this exercise were light commercial vehicles. The results from vehicle trials for light commercial vehicles show effective recognition of vehicle loading conditions with acceptable accuracy. The output of this novel method can be used for optimizing different ADAS functionalities at very low-cost leveraging telematics units.
Vaisakh Venugopal, Paul Raj Bob, Vipin Nair
Mahindra Research Valley
Symposium on International Automotive Technology 2019