A system, a method for training a machine learning based processor circuitry suitable for characterizing an envi-ronment of a vehicle

TL;DR: In this paper, a machine learning-based processor circuitry (114) was used for characterizing an environment of a vehicle (110) and a method (200) for training a machine-learning based processor circuitry(114) suitable for characterising an environment (110).

Abstract: Embodiments of the present disclosure relate to a system (100) for characterizing an environment of a vehicle (110) and a method (200) for training a machine learning based processor circuitry (114) suitable for characterizing an environment of a vehicle (110). The method (200) comprises obtaining first object information from first monitoring data generated by monitoring at least one object (130, 140, 160) within the environment with at least one first environmental sensor circuitry (120) statically installed within the environment. The method (200) further provides for generating second monitoring data by monitoring the object (130, 140, 160) with at least one second environmental sensor circuitry (112) installed at the vehicle (110), wherein the second monitoring data is time-synchronized with the first monitoring data. Further, the method (200) comprises modifying one or more data processing parameters of the machine learning based processor circuitry (114) based on the first object information and the second monitoring data to cause a deviation between the first object information and second object information obtained from the second monitoring data by using the machine learning based processor circuitry (114) to decrease.The system (100) comprises a processor circuitry (150) configured to obtain first object information from first monitoring data generated by monitoring at least one object (130, 140, 160) within the environment with at least one first environmental sensor circuitry (120) statically installed within the environment. The system (100) further comprises a second environmental sensor circuitry (112) installed at the vehicle (110) and configured to generate second monitoring data by monitoring the object (130, 140, 160). The second monitoring data is time-synchronized with the first monitoring data. Further, the system comprises a machine learning based processor circuitry (114). The machine learning based processor circuitry (114) is configured to modify one or more of its data processing parameters based on the first object information and the second monitoring data such that a deviation between the first object information and second object information obtained from the second monitoring data by using the machine learning based processor circuitry (114) decreases.