| In order to meet the environmental sensing requirements of autonomous vehicles,most existing autonomous vehicles are equipped with a variety of sensing systems such as millimeter wave radar,LIDAR and cameras.However,each sensing system is deployed independently and has different working mechanisms,making it difficult to coordinate control and interact effectively.Multi-systems such as on-board radar and communication are prone to co-channel and adjacent channel interference,which will seriously affect the performance of the sensing,communication and information processing systems of autonomous vehicles,which is a safety hazard for autonomous driving.Taking the on-mounted millimeter wave radar as an example,the problem of mutual interference among multiple radars operating in the same frequency band cannot yet be effectively resolved from the perspective of radar signal processing,leading to a significant reduction in the performance of radar detection distance and detection accuracy during the actual measurement of autonomous driving.At this stage,there are few research works that comprehensively analyze the influence of multi-radar mutual interference on detection performance in theory.Most of them adopt idealized radar interference modelling,which leads to unclear influence on the interference in actual road scenarios and a lack of accurate models and analysis methods that match the radar characteristics of actual scenarios.Accurate interference models and corresponding interference mitigation strategies are the cornerstone for improving networked sensing.Therefore,it is of great engineering value and urgent need to study interference analysis and interference mitigation strategies for networked autonomous driving.To address the problems of interference analysis and interference mitigation in networked autonomous driving scenarios,this paper conducts an extensive and in-depth study,with the following innovation points summarized.1.Modeling and Analysis of Radar Mutual Interference in Networked Autonomous Driving EnvironmentTo address the limitation that most of current research works on radar interference adopt idealized radar interference modelling methods,resulting in a lack of clarity on the extent of interference effects in actual road scenarios and a lack of accurate models and analysis methods that are consistent with the mutual interference among multiple radars.At this stage,there is little research work that comprehensively analyze the influence of multi-radar mutual interference on detection performance in theory.Therefore,this paper proposes a new mutual interference analysis framework,which not only uses stochastic geometry to quantify direct interference and reflected interference,and derives a closed-form solution,but also uses a deterministic analysis method to analyze adjacent channel interference to achieve modelling and analysis of radar mutual interference in actual road scenarios.2.Performance Analysis of Uncoordinated Interference Mitigation Approach for Automotive RadarTo address the lack of performance analysis of existing uncoordinated interference mitigation approaches and the requirement for higher detection performance,this paper evaluates the performance of two typical uncoordinated interference mitigation approaches,such as random frequency division multiplexing and frequency hopping.The performance in terms of the probability of false detection and missed detection,effective detectable density and maximum number of interference-free radar is analyzed precisely,and a closedform solution is obtained.Finally,to further improve the interference mitigation performance,this paper propose an innovative uncoordinated interference mitigation method by combining adaptive frequency hopping technique with binary phase modulation.Simulation results show that the uncoordinated interference mitigation approach can reduce the mutual interference probability to less than 10%.3.Research and Performance Analysis of Coordinated Interference Mitigation Approach for Automotive RadarTo address the lack of framework and performance analysis of coordinated interference mitigation approach,this paper proposes a framework for cellular based vehicle-to-everything communication-aided radar mutual interference mitigation and designs a time-frequency division based radar multi-access channel access scheme.Through communication coordination,multiple radars access the shared channel in a cooperative manner to avoid serious mutual interference among multiple radars and the interference mitigation approach is evaluated in terms of a number of performance metrics.Finally,a collaborative radar communication power allocation strategy is investigated to maximize the detection range of the system while ensuring the performance requirements of the radar and communication systems.Simulation results show that the coordinated interference mitigation approach can reduce the probability of interference to less than 0.9%.4.Research and Performance Analysis of Joint Communication and Sensing Interference Mitigation Approach for Automotive RadarIn response to the limited existing research work on radar mutual interference mitigation approach based on Joint Communication and Sensing(JCS),this paper proposes a new integrated multi-beam based JCS design scheme.Through the integrated multi-beam design,a constant gain fixed sub-beam is constructed using the transmitting antenna in the data transmission mode to perform the communication function,and a time-varying scanning sub-beam is constructed in the other directions to perform the sensing function.In addition,to address the lack of interference analysis framework under the JCS system,this paper propose a cross-system interference analysis framework and evaluate the performance of interference mitigation based on JCS.At the same time,since communication and sensing have different performance requirements,this paper investigate the joint resource allocation problem for the coexistence of multi-carrier communication and sensing systems.Simulation results show that the JCS based interference mitigation approach can reduce the probability of interference to less than 0.4%.Finally,we conclude the whole thesis and talk about the future work. |
PDF Full Text Request