10.12.2021 Sensor Fusion — It’s all about Prediction
Sensor fusion systems spend a significant amount of resources in predicting the future. Eric Richter explains how a multiple model approach can better predict different object classes and thus dramatically improve sensor fusion performance.
In summary, sensor fusion systems should support class-specific motion models using multiple model approaches to account for different object behaviors, to initialize tracks using multiple hypotheses to cope with initialization ambiguities,to apply different sensor models depending on the object class, and to efficiently handle the hypotheses to save CPU and memory resources.
Read Eric's article