EU_MAVEN D3.2: Cooperative environment perception algorithms

Kurzfassung

This document describes the project achievements in cooperative perception algorithm development and testing. An overview of the building blocks, vehicle setups and embeddings of the perception algorithms into the test vehicle architectures used within the experiments at DLR and HMETC is given. A review of the state-of-the-art outlines the different levels of sophistication that are known from literature for the association and data fusion steps of a cooperative environment perception algorithm and reflects recent developments in the field. Real time algorithms are explained that were used in the experiments and demonstrations at the HMETC test site with both HMETC and DLR vehicles involved. Fusion experiments demonstrate detection and perception of a pedestrian not in sight of the ego vehicle sensors via a collective perception (CPM) message and combining of position estimates received via a cooperative awareness message (CAM) with the sensor detections of a test vehicle. Based on the results of the state-of-the-art review, an advanced fusion algorithm is proposed and its performance in urban traffic scenarios as well as Monte Carlo simulation scenarios evaluated. Conclusions are drawn for cooperative perception algorithm development and testing. This deliverable serves (in combination with D3.3) as basis for the testing done in WP6 with its upcoming deliverable D6.4 and the evaluation done in WP7 with its upcoming deliverable D7.2. The developments, experiments and evaluations presented in this deliverable have been executed by the MAVEN project partners HMETC and DLR in tight cooperation.