2024 FFT

Forensic Trajectory Prediction

Chris Flake (MITRE)

Abstract: Commercial sensing systems are bringing mobility (and other) analytics to a wider audience than in the past. Previously, only governments could put high quality sensing systems into space. Now there are many options that are accessible to normal people and businesses. This has impacted multiple business sectors in the US including the financial industry for hedging investments. There are now emerging sensors that can be used to collect mobility data from space, which provides a rich dataset that could be further integrated into the business sector. Mobility data contains information corresponding to human intent, and therefore can be used beyond just image-based computer vision. But for these commercial sensors to be able to follow a single entity across multiple sensors, requires the ability to predict where objects will go into the future. This briefing goes through past efforts at implementing trajectory prediction, and concepts and metrics used in the testing of the prediction engine.