{"title":"Graph based Cooperation Strategies for Automated Vehicles in Mixed Traffic","authors":"Maximilian Flormann, Roman Henze","doi":"10.4271/2024-01-2982","DOIUrl":null,"url":null,"abstract":"In the context of urban smart mobility, vehicles have to communicate with each other, surrounding infrastructure, and other traffic participants. By using Vehicle2X communication, it is possible to exchange the vehicles’ position, driving dynamics data, or driving intention. This concept yields the use for cooperative driving in urban environments. Based on current V2X-communication standards, a methodology for cooperative driving of automated vehicles in mixed traffic scenarios is presented. Initially, all communication participants communicate their dynamic data and planned trajectory, based on which a prioritization is calculated. Therefore, a decentralized cooperation algorithm is introduced. The approach of this algorithm is that every traffic scenario is translatable to a directed graph, based in which a solution for the cooperation problem is computed via an optimization algorithm. This solution is either computed decentralized by various traffic participants, who share and compare their solutions in order to get an optimal one, or centralized by a single computation unit, such as smart infrastructure systems. The cooperation participants negotiate the cooperative driving maneuver via a chain like validation approach, since the communication protocol does not require any handshake by design. Finally, all cooperation participants carry out the optimized and negotiated cooperative driving maneuver. The presented algorithm is validated in a multi-vehicle simulation. Different optimization heuristics are compared, ranging from traditional approaches to machine learning algorithms. The methods' behavior with regard to increasing model complexities is evaluated based on a representative catalogue of scenarios. Finally, the algorithm is validated in a real world proving ground test. These validations show that the introduced methodology provides significantly more efficient cooperation strategies compared to traditional, infrastructure-controlled approaches. Additionally, the presented approach is conflict-free by design.","PeriodicalId":510086,"journal":{"name":"SAE Technical Paper Series","volume":"2 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"SAE Technical Paper Series","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.4271/2024-01-2982","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
In the context of urban smart mobility, vehicles have to communicate with each other, surrounding infrastructure, and other traffic participants. By using Vehicle2X communication, it is possible to exchange the vehicles’ position, driving dynamics data, or driving intention. This concept yields the use for cooperative driving in urban environments. Based on current V2X-communication standards, a methodology for cooperative driving of automated vehicles in mixed traffic scenarios is presented. Initially, all communication participants communicate their dynamic data and planned trajectory, based on which a prioritization is calculated. Therefore, a decentralized cooperation algorithm is introduced. The approach of this algorithm is that every traffic scenario is translatable to a directed graph, based in which a solution for the cooperation problem is computed via an optimization algorithm. This solution is either computed decentralized by various traffic participants, who share and compare their solutions in order to get an optimal one, or centralized by a single computation unit, such as smart infrastructure systems. The cooperation participants negotiate the cooperative driving maneuver via a chain like validation approach, since the communication protocol does not require any handshake by design. Finally, all cooperation participants carry out the optimized and negotiated cooperative driving maneuver. The presented algorithm is validated in a multi-vehicle simulation. Different optimization heuristics are compared, ranging from traditional approaches to machine learning algorithms. The methods' behavior with regard to increasing model complexities is evaluated based on a representative catalogue of scenarios. Finally, the algorithm is validated in a real world proving ground test. These validations show that the introduced methodology provides significantly more efficient cooperation strategies compared to traditional, infrastructure-controlled approaches. Additionally, the presented approach is conflict-free by design.