Research and/or engineering questions/objective Cooperative (or collective) perception enables connected and automated vehicles to wirelessly exchange information about the objects they perceive with their onboard sensors. It is thus conceived to increase the perception capabilities of connected and automated vehicles, improving their safety. However, the implementation of cooperative perception is challenged by the need to minimize the latency in the transmission and reception of information. In this context, the main objective of this paper is the implementation, validation and testing of a Collective Perception Service (CPS) following ETSI specifications. The proposed implementation is validated under varying conditions and its scalability is analyzed. Methodology The proposed CPS is implemented over Vanetza, which is an open-source implementation of the ETSI C-ITS protocol stack. The CPS architecture has been designed to comply with ETSI specifications while ensuring its scalability. The CPS generates CPMs (Collective Perception Messages) following the ASN1 format defined by ETSI. The CPMs are populated with information about detected objects and the onboard sensors following the ETSI CPM generation rules, that is obtained from an external module (e.g. the perception system of the vehicle). When a new CPM is generated, the corresponding protocol headers are appended, and the CPM is transmitted. On the receiver side, all the CPMs are processed and sent back to the perception system. Multiple tests have been designed for the functional validation of the implemented CPS and to evaluate its scalability. Each test emulates a scenario with one or more static or moving detected objects of different types. Results The tests have been used to perform the functional validation of the implemented CPS verifying the CPMs generated against the ETSI CPM generation rules. The CPMs generated during the tests have also been cross-validated against third parties (in particular against OSS ASN1 commercial development toolkit), demonstrating the correct implementation of the CPMs generated in ASN1 format. This cross validation ensures the compatibility of the implemented CPS. Finally, the tests have also been used to measure the latency introduced by the internal information processing of the CPS in different parts of the architecture. The results obtained show the impact of object quantity or message rate on the delay required to update the internal list of detected objects. They also quantify the time needed to generate a CPM depending on its content and therefore its scalability. Limitations of this study The implemented CPS has been extensively tested in this paper in a lab environment using off-the-shelf PCs. However, we are working on its adaptation and evolution to be used over an innovative automotive Telematics Control Unit from Idneo, using C-V2X and 5G New Radio release 15 as radio access technology. Also, the evaluation conducted in this paper focuses on the internal processing performed by the CPS for the management of the objects information and generation/reception of CPMs. Nevertheless, there are other components in the protocol stack that could also introduce additional latency and will be evaluated in our future work. What does the paper offer that is new in the field in comparison to other works of the author? This is a joint work between University Miguel Hernandez (UMH) and Idneo Technologies. UMH has extensively worked on the evaluation of the CPS defined by ETSI by means of simulation, and has contributed to its design as part of the ETSI standardization process. They are also experts on C-V2X (LTE-V2X and 5G NR V2X). Idneo includes in its portfolio high performance telematic control units for vehicles, motorcycles and e-bikes, including 5G New Radio, C-V2X and DSRC technologies. In this paper, we join forces/expertise for the implementation and evaluation of the first CPS following the recently published ETSI standard. Conclusion This paper implements, validates and tests a Collective Perception Service (CPS) for connected and automated vehicles following ETSI specifications. Extensive functional validation tests as well as cross-validation and scalability experiments have been conducted in a lab environment. The results obtained demonstrate its potential to be integrated in a telematic control unit for connected and automated vehicles. Our next steps include the adaptation of the implemented CPS and the C-ITS stack to work over C-V2X and its evaluation using a dedicated embedded hardware in real-world experiments as part of the InPercept project.

Mr. Santiago López, Software Engineer, Universidad Miguel Hernández (UMH)