Vehicle-mounted multi-source radar converged communication system based on FlexRay bus and synchronization method thereof

Abstract

The invention belongs to the technical field of radars, and discloses a vehicle-mounted multi-source radar converged communication system based on a FlexRay bus and a synchronization method of the vehicle-mounted multi-source radar converged communication system based on the FlexRay bus. The FlexRay communication bus supports a static and dynamic dual-channel structure and has a deterministic communication capability, and the maximum bandwidth can reach 10Mbps; the distributed radar controller is connected with each radar node and used for collecting local radar data, packaging timestamps and uploading the data through FlexRay; the time synchronization controller is used for realizing nanosecond synchronization through a FlexRay global clock mechanism and broadcasting clock signals to all radar nodes; the fusion processing unit is used for intensively receiving all radar data and carrying out target detection and data fusion based on a unified time sequence; and the main control coordination module is used for upwards communicating with the automatic driving domain controller and downwards uniformly coordinating the behaviors of the subsystems. The system has the capabilities of accurate time sequence, multi-source data fusion, fault-tolerant control and the like, and is a high-reliability communication platform supporting intelligent driving decision making.

Description

Technical Field

[0001] The present invention belongs to but is not limited to the field of radar technology, and in particular relates to a vehicle-mounted multi-source radar fusion communication system based on a FlexRay bus and a synchronization method thereof. Background Art

[0002] The current existing technology is a multi-source radar fusion system based on the CAN-FD communication bus. The system mainly connects the millimeter-wave radar and lidar nodes through the CAN-FD bus, and uploads the data to the central controller for processing. In actual deployment, CAN-FD has the advantages of low cost and mature interface, and is suitable for basic ADAS systems or low-speed automatic parking scenarios. However, due to the limited bandwidth of the CAN-FD bus (up to only 5Mbps) and the non-deterministic event triggering mechanism of the communication process, data congestion is very likely to occur when multiple radars upload at a high frame rate concurrently, resulting in uncontrollable transmission delay, which cannot meet the rigid requirements of L3 and above autonomous driving systems for high-frequency, multi-mode data synchronous fusion.

[0003] Furthermore, existing systems generally lack precise global clock synchronization mechanisms. Radar nodes rely on software timestamps or external trigger signals for approximate alignment, causing drift in the transmission paths and processing cycles between different radar data types. Timing errors can accumulate to millisecond levels. In scenarios with drastic target dynamics or high-speed driving, this timing mismatch can directly lead to misjudgment of fused targets or trajectory changes, reducing system perception accuracy and limiting its applicability in complex road conditions or medium- and high-speed autonomous driving missions. Summary of the Invention

[0004] In view of the problems existing in the prior art, the present invention provides a vehicle-mounted multi-source radar fusion communication system based on a FlexRay bus and a synchronization method thereof.

[0005] The present invention is implemented as follows: a vehicle-mounted multi-source radar fusion communication system based on a FlexRay bus, comprising:

[0006] Multi-source radar perception unit, FlexRay communication bus, distributed radar controller, time synchronization controller, fusion processing unit and main control coordination module;

[0007] The multi-source radar perception unit includes millimeter wave radar, laser radar and ultrasonic radar, which respectively collect different types of environmental perception data;

[0008] The distributed radar controller is connected one by one with the multi-source radar sensing units, and is used to add a timestamp to the sensing data and encapsulate and upload it;

[0009] The FlexRay communication bus adopts a dual-channel scheduling structure of static frame segments and dynamic frame segments to transmit the data uploaded by each radar controller to the fusion processing unit;

[0010] The time synchronization controller acts as a FlexRay master node and achieves global time alignment of each radar controller by periodically broadcasting synchronization frames;

[0011] The fusion processing unit is used to receive and fuse multi-source radar data and output a unified perception target;

[0012] The master control coordination module is used to manage the status of each radar node and upload the fusion results to the automatic driving control system.

[0013] Furthermore, each radar in the multi-source radar sensing unit integrates an analog-to-digital conversion module, a digital signal processing module and a data frame output module;

[0014] The distributed radar controller is configured with an SPI or CAN-FD interface for receiving radar perception data and completing timestamp embedding based on a local oscillator clock.

[0015] Furthermore, the static frame segments in the FlexRay communication bus are used to transmit time synchronization frames and high-priority perception data frames, and the dynamic frame segments are used to transmit non-critical control instructions;

[0016] The system adopts a fixed frame scheduling table to configure the frame transmission time slots of each radar controller to ensure the time domain determinism and inter-frame interference suppression during the static segment transmission process.

[0017] Furthermore, the time synchronization controller broadcasts a synchronization frame once in each communication cycle, and the radar controller corrects the local clock deviation according to the reference time value in the synchronization frame to achieve nanosecond-level synchronization accuracy.

[0018] Furthermore, the fusion processing unit includes a cache scheduling module, a time normalization module and a fusion reasoning module;

[0019] The time normalization module performs interpolation alignment according to the timestamps of the data frames;

[0020] The fusion reasoning module completes multi-source data fusion based on the Kalman filter algorithm and outputs the speed, position and category information of the target.

[0021] Furthermore, the master control coordination module is connected to the autonomous driving controller through the AUTOSAR communication interface, and sends the parameter configuration, sampling frequency and power consumption level of each radar controller through the CAN interface.

[0022] The present invention also provides a data processing method for a vehicle-mounted multi-source radar fusion communication system based on a FlexRay bus, comprising the following steps:

[0023] S1, radar collects raw data and completes analog-to-digital conversion;

[0024] S2, the distributed radar controller receives the raw data and binds the timestamp;

[0025] S3, encapsulated into FlexRay data frames and then uploaded to the bus according to the static or dynamic segment;

[0026] S4, the fusion processing unit receives the data frame and performs time normalization and target fusion;

[0027] S5, the fusion result is sent to the main control coordination module and the radar node operation status is fed back.

[0028] Furthermore, the data fusion includes:

[0029] Use millimeter wave radar data to obtain target speed information,

[0030] Reconstruct the target shape using lidar point cloud data,

[0031] Use ultrasonic data to compensate for blind spot information,

[0032] The unified target recognition is achieved through the Bayesian confidence fusion mechanism or JPDA algorithm.

[0033] Furthermore, the fusion processing unit uses a multi-source data buffer pool to receive each frame of data in parallel and performs alignment processing based on the timestamp;

[0034] Each data frame must pass a consistency check before entering the time normalization module. Data that exceeds the time threshold will be discarded and reported to the main control coordination module.

[0035] The present invention also provides a computer-readable storage medium, wherein the medium stores a computer program. When the program is executed by a processor, the processor executes the steps of the method according to any one of claims 7 to 9.

[0036] In combination with the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solutions to be protected by the present invention are as follows:

[0037] This invention effectively enhances the time-domain coordination capabilities of multi-source radar systems by introducing the FlexRay communication protocol and a distributed control architecture. Based on the FlexRay static master clock mechanism, the system synchronizes distributed radar controllers with nanosecond-level time accuracy through periodic broadcast of global synchronization frames and local oscillator phase calibration. Compared to the millisecond-level synchronization errors of the CAN bus or the significant clock drift of traditional Ethernet under high load, this system ensures that heterogeneous sensor data is aligned on the same time reference, improving the consistency and positioning accuracy of multi-sensor fusion from the source.

[0038] To address fusion processing latency, the system employs a static frame segment plus priority scheduling strategy at the bus layer, avoiding arbitration conflicts and frame retransmissions common in asynchronous communications. Key sensing nodes are pre-assigned fixed time slots to ensure that high-priority data is not blocked. The average latency of the fusion processing unit after receiving data is kept below 10ms. This response time is significantly superior to sensing systems based on CAN-FD or 100M Ethernet architectures, meeting the real-time requirements of a perception-planning closed-loop cycle of less than 20ms for Level 3 and higher autonomous driving systems.

[0039] The system's structural design supports flexible scalability and platform migration. The distributed radar controller, packaged based on a unified interface specification, supports seamless integration of multiple radar systems (such as FMCW millimeter-wave radar, MEMS lidar, and narrow-beam ultrasonic radar). The controller also provides a standardized data packaging interface and frame scheduling configuration table. This modular mechanism ensures the platform's rapid deployment and functional upgrade capabilities across different vehicle models and sensor topologies, aligning with the future trend of autonomous driving systems toward heterogeneous integration and customized perception layouts.

[0040] The FlexRay dual-channel communication mechanism provides the system with a technical foundation that balances high bandwidth and high reliability. Separate scheduling of static and dynamic segments ensures the parallel transmission of critical control instructions and large amounts of sensor data. If any channel fails, the system automatically switches to the redundant channel to ensure uninterrupted communication. Compared to traditional communication solutions that experience data frame congestion and frequent frame drops in complex environments, this system demonstrates extremely high data integrity and transmission stability, significantly enhancing the robustness of the vehicle's perception system in emergency scenarios (such as sudden obstacles and low-visibility environments).

[0041] The core advantage of this invention lies in its precise multimodal data timing mechanism and low-latency, high-confidence fusion processing capabilities. In a typical urban road test scenario, the system outputted target tracking trajectory error less than 8cm, and the fusion perception confidence level remained stable above 95%. The point cloud and radar reflection intensity seamlessly complement each other, significantly reducing recognition errors caused by occlusion, blind spots, or signal echo ambiguity, providing stable and reliable environmental input for subsequent path decision-making and regulatory control modules.

[0042] The system of this invention fully addresses the core technical requirements of autonomous driving scenarios: high precision, high real-time performance, high fault tolerance, and high scalability. Through the organic combination of FlexRay deterministic scheduling, time synchronization, and integrated intelligent control, this system not only builds a multi-source radar fusion communication hub for automotive-grade applications, but also provides critical infrastructure support for intelligent decision-making and safe operation in complex traffic environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 This is a structural diagram of a vehicle-mounted multi-source radar fusion communication system based on the FlexRay bus provided by an embodiment of the present invention;

[0044] Figure 2 is a flowchart of a detailed signal data processing process of the vehicle-mounted multi-source radar fusion communication system based on the FlexRay bus provided by an embodiment of the present invention;

[0045] Figure 3 This is a flow chart of a vehicle-mounted multi-source radar fusion communication method based on the FlexRay bus provided by an embodiment of the present invention;

[0046] Figure 4 This is an architecture diagram of a vehicle-mounted multi-source radar fusion communication system based on the FlexRay bus provided by an embodiment of the present invention;

[0047] In the figure: 1. Multi-source radar perception unit; 2. FlexRay communication bus; 3. Distributed radar controller; 4. Time synchronization controller; 5. Fusion processing unit; 6. Master control coordination module. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0049] like Figure 1 As shown, an embodiment of the present invention provides a vehicle-mounted multi-source radar fusion communication system based on a FlexRay bus, the system comprising:

[0050] Multi-source radar perception unit 1: includes millimeter-wave radar, lidar, ultrasonic radar, etc., each independently collecting environmental information;

[0051] FlexRay communication bus 2: supports static + dynamic dual-channel structure, has deterministic communication capabilities, and has a maximum bandwidth of up to 10Mbps;

[0052] Distributed radar controller 3: Connected to each radar node, it collects local radar data, packages it with timestamps, and uploads it via FlexRay;

[0053] Time synchronization controller 4: uses FlexRay's global clock mechanism to achieve nanosecond-level synchronization and broadcast the clock signal to all radar nodes;

[0054] Fusion processing unit 5: centrally receives all radar data and performs target detection and data fusion based on a unified time sequence;

[0055] Main control coordination module 6: Communicates with the autonomous driving domain controller upward and coordinates the behavior of each subsystem downward.

[0056] Current in-vehicle environmental perception systems face technical challenges such as heterogeneous multi-source radar deployment, high communication latency, and severe timing mismatches. Traditional CAN or Ethernet communication architectures, due to bandwidth and deterministic limitations, struggle to ensure synchronous data transmission across high-speed, multi-node radar systems. Low latency and high synchronization accuracy are crucial, especially in autonomous driving. This paper proposes a multi-source radar fusion communication system based on the FlexRay bus architecture, aiming to address key technical bottlenecks such as communication consistency, timing alignment, and low fusion processing accuracy across different radar devices.

[0057] First, the system deploys a variety of heterogeneous radar nodes, including millimeter-wave radar, lidar, and ultrasonic radar, to achieve independent perception of different environmental characteristics. Each of these radar nodes possesses local data collection and preliminary preprocessing capabilities, reducing the burden of redundant raw data transmission. However, due to the varying sensing frequencies and dispersed physical structures of heterogeneous radars, effective integration is difficult to achieve directly and requires the support of stable communication and synchronization mechanisms.

[0058] To this end, the FlexRay bus was introduced as the core communications foundation, leveraging its static frame segments to achieve highly deterministic, periodic broadcast data transmission. Dynamic frame segments, meanwhile, enable asynchronous bandwidth allocation based on mission urgency. The FlexRay bus inherently features dual-channel redundancy and bit-level synchronization, with a maximum bandwidth of 10 Mbps. This meets the millisecond-level high-frequency data upload requirements of multi-source radars, significantly improving communication real-time performance and system robustness.

[0059] Each distributed radar controller uses its local embedded processor to timestamp the raw radar data, encapsulates it in a frame structure, and uploads it to the FlexRay bus, standardizing the data structure on the acquisition side. This timestamp mechanism is tightly coupled with the bus's global synchronization mechanism, ensuring that data from different radar nodes has a unified reference time domain, avoiding target mismatches or trajectory drift caused by time differences during data fusion.

[0060] The time synchronization controller deployed in the system broadcasts synchronization frames based on FlexRay's time-triggered mechanism, achieving nanosecond-level time synchronization for all nodes within the system. By broadcasting global clock information, radar nodes share a unified reference timeline, ensuring strict consistency in data upload and fusion timing. As the clock master node, the synchronization controller maintains system clock stability, laying the foundation for orderly data collection and fusion.

[0061] Ultimately, all collected data is centrally sent to the fusion processing unit for unified processing. This unit performs spatial registration and feature redundancy elimination on a unified timeline for multi-radar target data, fusing them into high-confidence, multi-dimensional perception information. Simultaneously, the master control coordination module, acting as an upper-level interface, connects data with the autonomous driving domain controller to rapidly respond to fusion results and coordinates the status and operating strategies of each radar subsystem, building a complete, closed-loop on-board intelligent perception-communication system.

[0062] This implementation utilizes a test vehicle supporting Level 3 autonomous driving as the vehicle, building a multi-source radar fusion communication system based on the FlexRay bus. The front-end sensing configuration includes a 77GHz millimeter-wave radar installed on the front of the vehicle for mid- and long-range obstacle detection, a lidar on each wing for building a 3D point cloud model, and four ultrasonic radars on the lower edge of the vehicle for close-range blind spot detection. Each radar system is connected to an independent distributed radar controller for data collection and processing.

[0063] At the communication architecture level, the FlexRay bus structure adopts a star topology, with a dual-channel backbone for static and dynamic scheduling. Channel A is used for periodic transmission of time-aligned data frames across all nodes, while Channel B is used for asynchronous fault uploads or non-critical command communication, ensuring redundant fault tolerance in the event of a single-channel anomaly. A static time slot table is configured to assign fixed allocations to the radar controller for transmitting synchronously encapsulated data, ensuring deterministic time-domain scheduling and controllable delay bounds.

[0064] Each radar controller integrates a local sampling counter and a high-precision oscillator, which, combined with the global synchronization frame of the FlexRay master, performs periodic time synchronization. The controller encapsulates the raw radar detection data into time slices, calibrates the local timestamp, and attaches the node number before uploading it via the FlexRay bus to the centralized fusion processing unit. This time encapsulation rapidly maps the raw heterogeneous data into a unified global time domain, laying the foundation for subsequent data alignment and fusion.

[0065] The time synchronization controller operates in master-slave clock mode. The FlexRay master node sends a global clock synchronization frame at the beginning of each communication cycle. All slave radar controllers receive it and perform phase alignment and frequency calibration. This system achieves synchronization accuracy within ±100ns, effectively resolving target position drift caused by inconsistent sampling frequencies between lidar and millimeter-wave radar, and improving fusion accuracy and system consistency.

[0066] The fusion processing unit is hosted by a dedicated SOC hardware platform, featuring a built-in data timing scheduling engine and multi-source registration algorithm module. The system utilizes a multi-modal fusion framework combining Kalman filtering and point cloud spatial interpolation to achieve feature fusion, path continuity verification, and occlusion compensation for the same spatial target. The fusion results are mapped to the HD map reference coordinate system, and the global perception map is refreshed every 20ms to support real-time calls to the autonomous driving high-level decision-making modules.

[0067] The master control and coordination module operates based on the AUTOSAR architecture and features multi-task scheduling and communication interface management. It communicates with the vehicle domain controller via CAN, regularly issuing radar node activation configurations, parameter self-test instructions, and status recovery mechanisms. It also receives fusion results and abnormality warning data in real time, and drives driving strategy switching, speed regulation, and path reconstruction, achieving an integrated closed-loop perception-control system.

[0068] The integrated implementation of the multi-source radar perception unit 1 specifically includes:

[0069] The system deploys heterogeneous radar front-end nodes, such as 77GHz millimeter-wave radar for long-range target speed detection, MEMS lidar for building dense point clouds, and narrow-beam ultrasonic radar for near-field obstacle detection, which are independently mounted in different areas of the vehicle. These radar nodes all have built-in low-power MCUs and acquisition modules, which can collect raw data in real time at different frame rates, such as 20Hz to 100Hz, and perform a filtering and signal-to-noise ratio improvement on the signal, laying the preprocessing foundation for subsequent data transmission.

[0070] The frame-level scheduling strategy design of the FlexRay communication bus 2 specifically includes:

[0071] Based on the TDMA+FTDMA structure in the FlexRay protocol, the system divides the time period into static frame segments and dynamic frame segments. The static frame segments use fixed time slots to map high-priority radar synchronization frames, while the dynamic frame segments are used to transmit asynchronous perception data packets of each radar node. In view of the large amount of radar perception data and high frame rate, the system uses frame compression coding + frame segment scheduling table (FIB) mechanism to achieve communication load optimization and inter-frame interference suppression.

[0072] The protocol docking mechanism of the distributed radar controller 3 specifically includes:

[0073] Each distributed radar controller node has a standard interface with the mounted radar sensor, such as SPI, CAN-FD, LVDS, or EthernetPHY, and is equipped with a high-precision timing module, such as a TCXO / OCXO, for receiving the time reference broadcast by the time synchronization controller; the distributed radar controller has an embedded timestamp alignment mechanism to accurately bind the perception data to the global timestamp. After data packet encapsulation is processed according to the FlexRay network layer protocol, it is uploaded to the bus according to the frame table schedule.

[0074] The master-slave synchronization strategy of the time synchronization controller 4 specifically includes:

[0075] The time synchronization controller serves as the static master clock node of the FlexRay network and uses a distributed clock synchronization protocol to periodically broadcast synchronization frames. All distributed radar controllers act as slave nodes and perform time drift correction and instantaneous time reconstruction based on local clock deviations, achieving microsecond-level timing consistency and meeting the rigid requirements of time domain fusion between radars.

[0076] The heterogeneous data synchronization fusion mechanism of the fusion processing unit 5 specifically includes:

[0077] The fusion processing unit uses a multi-core embedded SoC as the processing core, equipped with a data reordering buffer pool and a timing normalization engine, which is used to align and interpolate asynchronously reported multi-source radar data; the fusion engine uses Bayesian filters or extended Kalman filters to achieve cross-correction and trajectory collaborative estimation of multiple radar targets, improving the robustness of spatial perception.

[0078] The upstream and downstream interactive control logic of the master control coordination module 6 specifically includes:

[0079] As the core controller of the vehicle's electronic and electrical architecture, the master control coordination module exchanges data with the fusion processing unit through the AUTOSARCP standard interface, and realizes high-frequency and low-latency communication with the autonomous driving domain controller through the DDS middleware; the master control coordination module issues task synchronization frames and power management instructions to each sensor subsystem to achieve a unified task beat, data sampling window and power supply strategy, thereby supporting the real-time closed-loop operation of the perception-planning-control link.

[0080] like Figure 2 As shown, an embodiment of the present invention provides a detailed signal data processing process based on the vehicle-mounted multi-source radar fusion communication system based on the FlexRay bus, which specifically includes:

[0081] S1: Radar raw signal acquisition and preliminary preprocessing:

[0082] Each radar sensor receives and amplifies the echo signal through the embedded analog front-end (AFE) module, and then converts the analog signal into a discrete digital signal using a high-speed analog-to-digital converter (ADC). Millimeter-wave radar generates a range-Doppler spectrum through FMCW demodulation, while lidar forms a dense point cloud matrix. Ultrasonic radar generates an intensity reflection map of close-range targets. Each type of signal undergoes noise reduction filtering, envelope extraction, preliminary target recognition, and frame packaging in the local DSP or MCU before being transmitted to the DRC.

[0083] S2: Timestamp Binding and Data Buffering:

[0084] After preliminary processing, the radar signal is sent to the DRC's input buffer queue. At this point, the global clock signal broadcast by the time synchronization controller is synchronized to the DRC's internal synchronization control unit. The system calculates relative delays using the local real-time clock (RTC) and performs hardware-level timestamp embedding, accurately binding the timestamp to the header of each frame of radar data. This time information is then used for data frame reordering and time-domain fusion to ensure consistency in spatial information between different radar sources.

[0085] S3: Data frame encapsulation and uplink scheduling:

[0086] Radar data with timestamps is encapsulated into static or dynamic frames according to the FlexRay network layer protocol. A CRC checksum mechanism is used to construct the frame header to ensure transmission reliability. Key control or synchronization frames are transmitted in fixed time slots in the static segment, while environmental perception data frames are scheduled and delivered in the dynamic segment. The DRC control module uses a network scheduling table and frame conflict detection logic to implement bus access arbitration between multiple nodes, ensuring low-latency and highly reliable data upload.

[0087] S4: Data buffering and normalization before central fusion processing:

[0088] After the FPCU receives data frames from the FlexRay bus, the DMA controller first distributes them to different source radar buffer pools. The time normalization engine then interpolates and aligns each frame based on the attached timestamps. This process ensures that all types of data are mapped to a unified spatial model using the same time base, thus resolving the issue of temporal drift during the fusion of asynchronous multi-source data.

[0089] S5: Data fusion and target recognition process:

[0090] Based on time series alignment, the system calls a multi-sensor data fusion module to estimate target states using a Bayesian confidence framework or an extended Kalman filter (EKF). Millimeter-wave data provides radial velocity and relative distance to the target, while lidar point clouds are used to construct spatial shape models, and ultrasonic radar provides blind spot compensation information. The fusion model uses data association methods such as the nearest neighbor method, JPDA, confidence weighting, and the Kalman prediction-update process to output a unified set of perceived targets, including a unified ID, position, velocity, and category.

[0091] S6: Fusion result output and link feedback mechanism:

[0092] The perception results are sent to the autonomous driving controller through the VCU for path planning and decision-making, and the fusion accuracy, confidence, and latency monitoring indicators are sent to the upper layer; at the same time, the VCU regularly sends sampling frequency adjustment instructions or power consumption level strategies to each radar node to form a closed-loop data-control coordination mechanism; this feedback mechanism realizes adaptive perception optimization and resource management of the system in different driving scenarios by adjusting the radar task load and working status.

[0093] like Figure 3 As shown, an embodiment of the present invention provides a FlexRay bus-based vehicle-mounted multi-source radar fusion communication method based on the FlexRay bus-based vehicle-mounted multi-source radar fusion communication system, the method specifically comprising:

[0094] S21: The system initializes all radar nodes to register in the FlexRay network, the master node initializes the synchronous clock, and the time reference is broadcast to the entire network;

[0095] S22: Radar data collection and time calibration: Each radar node collects data and marks the timestamp with the received global synchronization signal;

[0096] S23: Synchronous data packaging and transmission, DRC encapsulates radar data and synchronization timestamps into FlexRay data frames and sends them to the bus in static frame slots;

[0097] S24: Central fusion unit: The FPCU receives radar data uploaded by each node and performs spatial consistency fusion based on time alignment to form a multi-dimensional perception map.

[0098] S25: Upper-level behavioral decision output, submitting the fused multi-source data to the vehicle's autonomous driving system for tasks such as path planning, collision warning, and target recognition.

[0099] In this implementation, the system first completes the initial registration of radar nodes within the FlexRay network. The master controller maps each radar module into the static time slot table within the FlexRay network according to pre-set addresses and communication parameters. Simultaneously, the FlexRay master node initiates a global synchronization mechanism, periodically broadcasting time synchronization frames. After receiving the initial synchronization frame, each distributed radar controller initiates local oscillator phase lock and calibrates the RTC to achieve a unified time base alignment across all nodes within the system within nanoseconds.

[0100] Subsequently, the millimeter-wave radar, lidar, and ultrasonic radar, each in operation, perform environmental perception tasks at its own sampling frequency. During data acquisition, each controller uses its local time management unit to embed a synchronized timestamp in each frame of data, completing data calibration. To ensure consistency during asynchronous data uploads, the controller implements a first-in, first-out structure in the cache management area to organize the sampled data. This allows data uploaded by different radar types at different cycles to be synchronously encapsulated on the bus using standard protocol frames.

[0101] The FlexRay communication bus plays a crucial role in data transmission within this architecture. The controller packages timestamped data into FlexRay static segment data frames, performs CRC integrity checks, and then uploads them to the backbone communication network in assigned static time slots. The dynamic segment is used to transmit status signals or non-critical data, avoiding congestion in the transmission of core sensing tasks. The frame scheduler is regularly refreshed under the control of the master node to ensure that data from all nodes is effectively reported within the fault-tolerance window.

[0102] The FPCU handles the central processing. Upon receiving data frames, its internal frame buffer scheduling module normalizes the data based on timestamps and node numbers. The fusion engine module integrates the spatial information provided by different radar types using a three-stage process: target prediction, matching, and updating. LiDAR describes the target's outline and voxel structure, while millimeter-wave radar provides velocity and radial accuracy data. Ultrasonic radar completes blind spot boundaries, ultimately outputting a multidimensional perception map.

[0103] The generated perception results are output to the master coordination module in the form of a multi-target status list. The module houses a protocol conversion and behavioral interface encapsulation unit, responsible for converting perceived target information into a data format compatible with the autonomous driving domain controller. Based on the results, the module dynamically updates control strategies, such as speed regulation, obstacle avoidance command triggering, and path reconstruction. Simultaneously, the module sends parameter scheduling commands back to each radar controller to adjust power consumption levels, switch frame rates, and manage node status, achieving a closed-loop feedback loop across the communication, perception, and control domains.

[0104] The system has demonstrated stable fusion response and high robustness in actual deployment. In real-world measurements in simulated urban scenarios, the FlexRay solution maintained multi-source fusion error within ±8 centimeters, maintained a data fusion cycle within 20 milliseconds, and supported up to six nodes concurrently uploading data, improving overall communication stability by approximately 40%. This solution establishes a highly real-time, scalable, and low-latency perception and communication infrastructure for mid- to high-level autonomous driving platforms.

[0105] An embodiment of the present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the vehicle-mounted multi-source radar fusion communication method based on a FlexRay bus.

[0106] The FlexRay bus-based on-board multi-source radar fusion communication system described in this invention is primarily used in advanced driver assistance systems (ADAS) and L3 and above autonomous driving systems. It is particularly suitable for passenger cars, smart commercial vehicles, unmanned delivery vehicles, and specialized autonomous driving platforms that require multimodal environmental perception and high-precision target fusion positioning. Furthermore, the system can also be expanded to roadside units in smart public transportation and vehicle-road collaborative V2X systems. As a core component for high-temporal consistency data scheduling in heterogeneous sensor networks, it is compatible with current mainstream autonomous driving electronic and electrical architectures (such as centralized EEA or regional controller architectures) and can seamlessly interface with the master domain controller.

[0107] Related products include:

[0108] 1. Vehicle-mounted distributed radar data acquisition module (DRC module) supporting FlexRay protocol;

[0109] 2. Edge computing fusion unit (FPCU) that integrates fusion algorithms and FlexRay parsing capabilities;

[0110] 3. Clock master controller with high-precision oscillator and synchronization scheduling mechanism;

[0111] 4. Software middleware control system that can be embedded in AUTOSAR;

[0112] 5. Automotive-grade perception kit assembly supporting multiple radar interfaces, targeted at OEMs and Tier-1 suppliers.

[0113] To verify the convergence consistency and communication performance of the proposed system, comparative experiments were conducted in closed scenarios and under real-world road conditions. Test vehicles were equipped with both the FlexRay bus solution and the CAN-FD architecture, each equipped with four 77GHz millimeter-wave radars, two LiDARs, and six ultrasonic radars, while operating an equivalent set of perception tasks. After 12 hours of data collection and analysis across various scenarios, the FlexRay solution demonstrated an average latency of 1.27ms for each radar packet, with maximum frame jitter kept within 0.21ms. However, the CAN-FD architecture exhibited latency fluctuations of up to 4.9ms, with some data experiencing congestion and packet loss.

[0114] In timing synchronization verification, based on the broadcast mechanism of the FlexRay master-slave clock protocol, radar node time drift fluctuated within 100ns, far superior to traditional trigger-based synchronization methods (fluctuations range from 3 to 5μs). Through synchronous timing record analysis, different radar data can be stably mapped to the same time slice, achieving cross-modal spatial alignment accuracy of less than 10cm during the fusion phase, improving fusion accuracy by approximately 23.5% compared to Ethernet asynchronous systems.

[0115] The effectiveness of fusion processing is quantified based on the onboard perception target consistency metric. In typical intersection, mixed urban driving conditions, and high-speed lane change scenarios, the FlexRay system maintained a 97.3% success rate for multi-source radar detection of the same target, compared to 88.9% for the traditional system. Fusion stability remained above 95% under high-speed driving conditions (>90 km / h), demonstrating the solution's suitability for the multi-radar perception requirements of medium- and high-speed autonomous driving scenarios.

[0116] At the same time, the operating data of the FPCU fusion processor based on SoC hardware shows that the system can complete the normalization, alignment and target estimation processing of 320 frames of data per cycle on average, and the overall processing delay is maintained between 18\ and 22ms, meeting the 20ms-level perception closed-loop response time requirement for autonomous driving; and the multi-sensor redundant processing logic effectively reduces target loss due to obstruction or failure of a single radar, enhancing the robustness of the system in complex traffic environments.

[0117] To further evaluate energy efficiency and scheduling controllability, the power consumption changes of each radar node were monitored during the test. Combined with the dynamic scheduling strategy of the main control coordination module, the system was able to reduce average power consumption by 18% by lowering the operating frequency of the ultrasonic radar when the vehicle speed was below 15 km / h. After adopting the low frame rate sampling strategy of millimeter-wave radar in sparse target areas, the communication load was reduced by 22% and the processing load was reduced by 17%, significantly improving the efficiency of system resource utilization.

[0118] Through actual vehicle deployment, heterogeneous network comparison testing and multi-scenario fusion consistency analysis, it is fully verified that the embodiments of the present invention have significant advantages in communication real-time, synchronization accuracy, target fusion quality and system resource scheduling, providing key technical support for high-level autonomous driving environment perception platforms.

[0119] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware portion can be implemented using dedicated logic; the software portion can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. Those skilled in the art will appreciate that the above-mentioned devices and methods can be implemented using computer-executable instructions and / or contained in processor control code, for example, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The devices and modules of the present invention can be implemented by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., can also be implemented by software executed by various types of processors, or can be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.

[0120] The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with this technical field within the technical scope disclosed by the present invention and within the spirit and principles of the present invention should be covered by the scope of protection of the present invention.

Claims

1. A vehicle-mounted multi-source radar fusion communication system based on FlexRay bus, characterized in that: include: Multi-source radar perception unit, FlexRay communication bus, distributed radar controller, time synchronization controller, fusion processing unit and main control coordination module; The multi-source radar perception unit includes millimeter wave radar, laser radar and ultrasonic radar, which respectively collect different types of environmental perception data; The distributed radar controller is connected one by one with the multi-source radar sensing units, and is used to add a timestamp to the sensing data and encapsulate and upload it; The FlexRay communication bus adopts a dual-channel scheduling structure of static frame segments and dynamic frame segments to transmit the data uploaded by each radar controller to the fusion processing unit; The time synchronization controller acts as a FlexRay master node and achieves global time alignment of each radar controller by periodically broadcasting synchronization frames; The fusion processing unit is used to receive and fuse multi-source radar data and output a unified perception target; The master control coordination module is used to manage the status of each radar node and upload the fusion results to the automatic driving control system.

2. The system according to claim 1, wherein: Each radar in the multi-source radar sensing unit is integrated with an analog-to-digital conversion module, a digital signal processing module and a data frame output module; The distributed radar controller is configured with an SPI or CAN-FD interface for receiving radar perception data and completing timestamp embedding based on a local oscillator clock.

3. The system according to claim 1, wherein: The static frame segment in the FlexRay communication bus is used to transmit time synchronization frames and high-priority perception data frames, and the dynamic frame segment is used to transmit non-critical control instructions; The system adopts a fixed frame scheduling table to configure the frame transmission time slots of each radar controller to ensure the time domain determinism and inter-frame interference suppression during the static segment transmission process.

4. The system according to claim 1, wherein: The time synchronization controller broadcasts a synchronization frame once in each communication cycle, and the radar controller corrects the local clock deviation according to the reference time value in the synchronization frame to achieve nanosecond-level synchronization accuracy.

5. The system according to claim 1, wherein: The fusion processing unit includes a cache scheduling module, a time normalization module and a fusion reasoning module; The time normalization module performs interpolation alignment according to the timestamps of the data frames; The fusion reasoning module completes multi-source data fusion based on the Kalman filter algorithm and outputs the speed, position and category information of the target.

6. The system according to claim 1, wherein: The master control coordination module is connected to the autonomous driving controller through the AUTOSAR communication interface, and sends the parameter configuration, sampling frequency and power consumption level of each radar controller through the CAN interface.

7. A data processing method for a vehicle-mounted multi-source radar fusion communication system based on FlexRay bus, characterized in that: The following steps are involved: S1, radar collects raw data and completes analog-to-digital conversion; S2, the distributed radar controller receives the raw data and binds the timestamp; S3, encapsulated into FlexRay data frames and then uploaded to the bus according to the static or dynamic segment; S4, the fusion processing unit receives the data frame and performs time normalization and target fusion; S5, the fusion result is sent to the main control coordination module and the radar node operation status is fed back.

8. The method according to claim 7, characterized in that The data fusion includes: Use millimeter wave radar data to obtain target speed information, Reconstruct the target shape using lidar point cloud data, Use ultrasonic data to compensate for blind spot information, The unified target recognition is achieved through the Bayesian confidence fusion mechanism or JPDA algorithm.

9. The method according to claim 7, characterized in that The fusion processing unit uses a multi-source data buffer pool to receive each frame of data in parallel and performs alignment processing based on the timestamp; Each data frame must pass a consistency check before entering the time normalization module. Data that exceeds the time threshold will be discarded and reported to the main control coordination module.

10. A computer-readable storage medium, wherein a computer program is stored in the medium, and when the program is executed by a processor, the processor is caused to perform the steps of the method according to any one of claims 7 to 9.

Images