Fuel is the biggest operating expense in the logistics industry, making up for 40-45% of the overall cost. Fuel pilferage remains a stifling issue that logistics providers have been grappling with, mainly because of unavailability of reliable and cutting-edge technology to accurately detect this on a real-time basis. More than 50% of fuel pilferages happen in moving trucks. Despite some of the logistics providers investing in IoT, there has been little impact, majorly due to the inability to handle fluctuations in data and low accuracy of sensors.

At Rivigo, we have built core technologies to solve this problem. Our team has worked relentlessly with our partners on their backend technologies as well as conducted field experiments to ensure that our sensors are ~100% accurate. We have developed in-house diagnostic engines and dashboards to detect fuel pilferage with ~100% accuracy.

We have used smoothening algorithms to remove fluctuations from the raw data, while still being able to detect even the slightest drop in the fuel level. Using advanced mathematical and statistical modelling, the results of these algorithms have been tested through extensive set of on-field trials and experiments. In some cases, the team at Rivigo Labs simulated truck fuel pilferages themselves!

In addition to pilferage detection, it is also critical to measure the filling efficiency of the fuel stations. Statistical modelling and geo-fencing techniques have been used to track all the fillings in our vehicles. The start and end values of the fuel are computed using local averages for better visibility. Fuel efficiency is also largely dependent on the driver’s driving behaviour. The most efficient drivers get about 30% better fuel efficiency than the least efficient ones. To improve the driving behaviour of our pilots, we have identified different sets of parameters like average speed, vehicle idling, over speeding and speed volatility to track fuel efficiency and have run different models like logistical regression and random forest to validate the same. We are now developing a fuel bot to identify driver-wise problems and recommend apt solutions.

Using advanced technology systems, we have been able to deliver impact not just by enabling huge cost savings but more importantly by changing driver behaviour to prevent pilferages and incentivising them to deliver better fuel efficiency.