Privacy Enhancement for Delivery Route Optimization through Occupancy Prediction

Delivery route optimization for logistic industry is one of applications proposed on smart meter infrastructure. It's expected to drastically reduce absent-delivery which amounts to 20% of total delivery in Japan, which estimated to save $billions a year. But as previous works pointed out, the concern on user privacy is the biggest hurdle yet to be addressed. In this research, we proposed a new approach to improve user privacy by converting electricity data into route data and optimize it before providing to service provider. Then, we tested pragmatic privacy improvement and route optimization through actual delivery experiment. Results showed that the information leakage rate (# of absence detection per delivery) decreased from 23% to 4% by this system and decreased to 2% with additional operational change. Also, the experiment validated decrease of absent-delivery rate from 23% to 2% and travel distance by 5% while improving privacy. Applying adequate method to "delivery optimization through occupancy prediction" enabled achieving both user privacy and absent-delivery reduction significantly.