APT3: Automated Product Traceability Trees Generated from GPS Tracks

With increasing concerns about food safety in many countries, product traceability has become an important risk-management tool. It enables the identification of possible sources of defective goods and facilitates the withdrawal and recall of affected products to protect consumers from foodborne diseases. However, it is troublesome for farmers to maintain the records required by high precision product fraceability during harvesting because traditional traceability systems usually involve human labor in paperwork or expenses for equipment purchase and installation; in either case, the resulting records are tedious. In this article, a fully automatic algorithm is proposed for efficiently generating product fraceability trees to visualize and store the full transport record of wheat from fields to elevators. Extending previous work on harvesting activity recognition via GPS tracks, this algorithm powers our fully automatic prototype product traceability system, APT3, and demonstrates great potential for tracing products solely with the GPS logs of the vehicles involved in harvest and transport. From the output traceability trees, the product yielded at any point in the field can be tracked all the way to the elevator where it was sold or stored by starting from the corresponding leaf node in the fraceability free and walking to the root of the free. Furthermore, each truckload of product unloaded at any destination elevator can be traced back to where the product was harvested by following the tree in the opposite direction. In this way, the fraceability records can be clearly visualized for farmers and easily used by other algorithms.