Automatic driving path planning and dynamic obstacle avoidance method for electric vehicle

Abstract

The invention discloses an automatic driving path planning and dynamic obstacle avoidance method for an electric vehicle, particularly relates to the technical field of vehicle control systems, and is used for solving the problem of control instability caused by the fact that real-time performance degradation of a driving system is not considered during path planning of an existing automatic driving system. A historical load spectrum and a real-time operation condition of a driving system are acquired in real time, a performance attenuation state is judged based on frequency domain feature conformity, thermal management strategy guidance is generated in combination with thermal inertia characteristics and obstacle avoidance urgency, and an action feasibility space meeting internal dynamic constraint and external space constraint at the same time is constructed. A system performance constraint evaluation result is obtained by analyzing a dynamic envelope, a real-time maximum output capability boundary of a driving system is predicted, a dynamic obstacle avoidance path is re-planned based on the boundary, and finally, the vehicle is controlled to execute obstacle avoidance operation to realize dynamic matching between the physical execution capability of the vehicle and a path planning requirement. And the performability of the obstacle avoidance path in the actual capability range of the vehicle is ensured.

Description

Technical Field

[0001] This invention relates to the field of vehicle control system technology, and more specifically, to an autonomous driving path planning and dynamic obstacle avoidance method for electric vehicles. Background Technology

[0002] Existing electric vehicle autonomous driving systems typically generate trajectories and make control decisions based on the vehicle's performance model under ideal conditions during path planning and dynamic obstacle avoidance. They identify dynamic obstacles through environmental perception modules and plan obstacle avoidance paths according to preset vehicle dynamic parameters. Finally, they complete the tracking operation through drive-by-wire actuators. Such methods generally assume that the drive system, steering system, and braking system can continuously provide the power output and response performance required for planning.

[0003] However, in actual operation, especially under continuous or extreme obstacle avoidance conditions, the electric drive system of electric vehicles is prone to heat accumulation due to high power output, which may lead to the degradation of system output performance. As a result, the dynamic obstacle avoidance path planned based on the ideal performance model cannot be effectively tracked, and there is a risk of vehicle control instability and obstacle avoidance failure. Existing methods lack a dynamic perception and fusion mechanism for the boundary of the vehicle's real-time physical execution capability, making it difficult to guarantee the feasibility and safety of path planning results under all operating conditions. Summary of the Invention

[0004] In order to overcome the above-mentioned deficiencies of the prior art, the present invention provides an autonomous driving path planning and dynamic obstacle avoidance method for electric vehicles to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: Autonomous driving path planning and dynamic obstacle avoidance methods for electric vehicles include: S1. Real-time acquisition of historical load spectrum, real-time operating conditions and vehicle operating environment data of electric vehicle drive system; S2. Based on the frequency domain characteristics of the historical load spectrum and the real-time operating conditions, determine whether the drive system is in or about to enter a performance degradation state. S3. When the system is in or about to enter a state of performance degradation, a time-domain trade-off evaluation is performed based on the thermal inertia characteristics of the drive system and the urgency of obstacle avoidance actions to generate a thermal management strategy guide. S4. Based on the thermal management strategy and vehicle operating environment data, construct an action feasibility space that simultaneously satisfies internal dynamic constraints and external spatial constraints, and obtain the system performance constraint evaluation results by analyzing its dynamic envelope. S5. Predict the real-time maximum output capability boundary of the drive system based on the system performance constraint evaluation results. S6. Based on the real-time maximum output capacity boundary and vehicle operating environment data, replan the dynamic obstacle avoidance path and control the electric vehicle to perform obstacle avoidance operations.

[0006] Furthermore, real-time acquisition of historical load spectra, real-time operating conditions, and vehicle operating environment data of the electric vehicle drive system, including: Obtain pre-stored load spectrum data of the drive system within historical operating cycles; The torque command, actual output torque, and speed of the drive motor at the current moment are collected as real-time operating conditions. The vehicle receives information on the location, speed, and road geometry of surrounding obstacles from environmental perception sensors as data for its operating environment.

[0007] Furthermore, based on the consistency between the historical load spectrum and the frequency domain characteristics of real-time operating conditions, it is determined whether the drive system is in or about to enter a performance degradation state, including: The historical load spectrum and the real-time operating conditions are transformed in the frequency domain to obtain the corresponding historical frequency domain characteristics and real-time frequency domain characteristics. The similarity of the distribution of historical frequency domain features and real-time frequency domain features in key frequency bands is calculated as the frequency domain feature conformity: the power spectral density distribution of historical and real-time operating condition signals in the main energy frequency band is extracted; the correlation coefficient of the power spectral density distribution of historical and real-time operating condition signals in the same key frequency band is calculated; and the correlation coefficient of the power spectral density distribution is used as a quantitative indicator of the frequency domain feature conformity. If the frequency domain feature compliance is lower than the preset compliance threshold, the drive system is determined to be in or about to enter a performance degradation state.

[0008] Furthermore, when the system is in or about to enter a performance degradation state, a time-domain trade-off assessment is performed based on the thermal inertia characteristics of the drive system and the urgency of obstacle avoidance actions to generate a thermal management strategy guide, including: The thermal relaxation time of the drive system is estimated based on its thermal capacity and thermal resistance parameters to characterize its thermal inertia properties. The obstacle avoidance action time window is calculated based on the relative motion relationship of dynamic obstacles in the vehicle operating environment data to characterize the urgency of the obstacle avoidance action; The ratio of thermal relaxation time to obstacle avoidance action time window is used as a time-domain trade-off factor. Based on the continuous numerical range of the time-domain trade-off factor, a thermal management strategy guide that smoothly transitions between thermal protection and obstacle avoidance performance is dynamically generated.

[0009] Furthermore, the smaller the time-domain trade-off factor, the more the thermal management strategy tends to prioritize thermal protection; the larger the time-domain trade-off factor, the more the thermal management strategy tends to prioritize obstacle avoidance performance.

[0010] Furthermore, based on thermal management strategy guidance and vehicle operating environment data, an action feasibility space that simultaneously satisfies internal dynamic constraints and external spatial constraints is constructed. The system performance constraint evaluation results are obtained by analyzing its dynamic envelope, including: The thermal management strategy is mapped to a range of constraints on the vehicle's longitudinal and lateral accelerations to form internal dynamic constraints; The geometric boundaries of the vehicle's drivable area are generated based on the obstacle locations, road boundaries, and traffic rules in the vehicle's operating environment data to form external spatial constraints. Establish a multi-constraint optimization problem with internal dynamic constraints as inequality constraints and external space constraints as equality constraints; By iteratively solving the feasible solution set of the multi-constraint optimization problem in real time, a feasibility space for action is constructed that is updated in real time with the movement of the vehicle. The coupling relationship between the change in the curvature of the envelope boundary and the shrinkage rate of the envelope volume in the prediction time domain of the action feasibility space was analyzed. Based on the synergistic evolution pattern of envelope boundary curvature change and envelope volume shrinkage rate, the system performance constraint evaluation results are determined.

[0011] Furthermore, by iteratively solving the feasible solution set of the multi-constraint optimization problem in real time, a real-time action feasibility space is constructed as the vehicle moves. This includes: initializing the current vehicle state initial value of the multi-constraint optimization problem in each control cycle; using numerical optimization algorithms to solve the vehicle state set that satisfies the internal dynamic inequality constraints and the external spatial equality constraints; and using the obtained feasible solution set as the action feasibility space at the current moment.

[0012] Furthermore, based on the co-evolutionary pattern of envelope boundary curvature change and envelope volume shrinkage rate, the system performance constraint evaluation results are determined, including: monitoring time-series data of envelope boundary curvature change rate and envelope volume shrinkage rate; establishing a time-domain correlation model between envelope boundary curvature change rate and envelope volume shrinkage rate; and generating an evaluation result of intensified system performance constraints when a co-deterioration pattern of increased curvature change and accelerated volume shrinkage is detected.

[0013] Furthermore, based on the system performance constraint evaluation results, the real-time maximum output capability boundary of the drive system is predicted, including: Extract the changes in envelope boundary curvature and envelope volume shrinkage rate from the system performance constraint evaluation results; Map the curvature variation of the envelope boundary to the maximum torque limit for the sustainable output of the drive system; Map the envelope volume shrinkage rate to the maximum allowable rate of change of rotational speed of the drive system; By combining the maximum torque limit and the maximum speed change rate limit, the real-time maximum output capability boundary of the drive system in the prediction time domain is generated.

[0014] Furthermore, based on the real-time maximum output capacity boundary and vehicle operating environment data, the dynamic obstacle avoidance path is replanned, and the electric vehicle is controlled to perform obstacle avoidance operations, including: A speed profile that conforms to the instantaneous power limit of the drive system is generated based on the real-time maximum output capability boundary. By combining obstacle location and road geometry information from vehicle operating environment data, an obstacle avoidance trajectory that satisfies vehicle kinematics is planned under velocity profile constraints. The obstacle avoidance trajectory is converted into drive motor torque commands and steering system angular displacement commands; The vehicle bus sends corresponding commands to the drive motor controller and steering system controller to perform obstacle avoidance operations.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. By real-time monitoring and performance prediction of the drive system's operating status, accurate perception of the vehicle's actual execution capability is achieved. By analyzing the frequency domain characteristics of historical load spectrum and real-time operating conditions, the drive system's performance degradation trend can be identified in advance. Combined with thermal inertia characteristics and obstacle avoidance urgency, a time-domain trade-off assessment is performed to generate an adaptive thermal management strategy. Based on the dynamic perception mechanism of physical execution capability boundaries, control instability caused by system performance degradation is effectively avoided.

[0016] 2. By constructing an action feasibility space that simultaneously satisfies internal dynamic constraints and external spatial constraints, dynamic matching between vehicle execution capabilities and path planning requirements is achieved. Based on the real-time maximum output capability boundary predicted by the system performance constraint evaluation results, accurate physical constraints are provided for path planning, thereby ensuring that the generated obstacle avoidance path is executable within the current actual capability range of the vehicle, significantly improving the safety and reliability of the autonomous driving system under complex working conditions. Attached Figure Description

[0017] Figure 1 This is a flowchart of the autonomous driving path planning and dynamic obstacle avoidance method for electric vehicles according to the present invention. Detailed Implementation

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0019] Example: Figure 1 This invention presents an autonomous driving path planning and dynamic obstacle avoidance method for electric vehicles, comprising: S1. Real-time acquisition of historical load spectrum, real-time operating conditions and vehicle operating environment data of electric vehicle drive system; S2. Based on the frequency domain characteristics of the historical load spectrum and the real-time operating conditions, determine whether the drive system is in or about to enter a performance degradation state. S3. When the system is in or about to enter a state of performance degradation, a time-domain trade-off evaluation is performed based on the thermal inertia characteristics of the drive system and the urgency of obstacle avoidance actions to generate a thermal management strategy guide. S4. Based on the thermal management strategy and vehicle operating environment data, construct an action feasibility space that simultaneously satisfies internal dynamic constraints and external spatial constraints, and obtain the system performance constraint evaluation results by analyzing its dynamic envelope. S5. Predict the real-time maximum output capability boundary of the drive system based on the system performance constraint evaluation results. S6. Based on the real-time maximum output capacity boundary and vehicle operating environment data, replan the dynamic obstacle avoidance path and control the electric vehicle to perform obstacle avoidance operations.

[0020] In the process of acquiring historical load spectrum, real-time operating conditions, and vehicle operating environment data of the electric vehicle drive system in real time, the first step is to acquire pre-stored load spectrum data of the drive system within the historical operating cycle. The historical operating cycle typically refers to the data period recorded during a continuous period of vehicle operation, such as the past 7 days or 30 days. The load spectrum data specifically refers to the statistical distribution of a series of load states experienced by the drive system within this historical operating cycle, recording the continuous operating time and frequency of the drive motor under different torque and speed combinations. This data is usually pre-stored in the vehicle's non-volatile memory, such as the EEPROM or Flash memory of the electronic control unit, in the form of a two-dimensional matrix or data list.

[0021] The stored load spectrum data includes, but is not limited to, the output torque value of the drive motor, the corresponding speed value, and the cumulative running time at each operating point. The torque value is typically measured in Newton-meters (N·m), and the speed value is typically measured in revolutions per minute (r / min). In actual acquisition, this historical data is read by accessing a specified memory address or calling a predefined data interface function, providing a data foundation for subsequent frequency domain analysis. The reading process includes data verification steps, such as using a cyclic redundancy check (CRC) method to verify data integrity and ensure the accuracy and reliability of the acquired data.

[0022] Next, the torque command, actual output torque, and speed of the drive motor at the current moment are collected as real-time operating conditions. The torque command refers to the target torque value issued by the vehicle controller to the drive motor, typically transmitted as a digital signal via the Controller Area Network (CAN) bus. The actual output torque refers to the actual torque value generated by the drive motor, obtained through the torque estimation algorithm built into the motor controller or by sensor measurement. The torque estimation algorithm is usually calculated based on the motor phase current and rotor position information. The speed refers to the actual rotational speed of the drive motor rotor, typically measured by an encoder or resolver. The acquisition process is achieved by directly reading the corresponding message data from the vehicle network bus, such as parsing the torque command value from the CAN bus, and simultaneously reading the real-time values ​​of the actual output torque and speed from the motor controller. This real-time data is collected at a fixed sampling period, typically 10 milliseconds to 100 milliseconds, ensuring that key characteristics of the drive system's dynamic response are captured. The collected data undergoes preprocessing, including unit standardization and data validity verification. For example, torque values ​​are converted to Newton-meters (N·m) in the International System of Units (SI), and speed values ​​are converted to revolutions per minute (r / min). Abnormal data points that are clearly beyond the physical possible range are removed.

[0023] Finally, the vehicle receives the surrounding obstacle positions, speeds, and road geometry information output from the environmental perception sensors as its operating environment data. Environmental perception sensors include, but are not limited to, millimeter-wave radar, lidar, vision cameras, and positioning systems. The position of surrounding obstacles refers to the coordinates of surrounding vehicles, pedestrians, and other obstacles relative to the vehicle, detected by the sensors. This is typically expressed in a Cartesian or polar coordinate system, with the coordinate values ​​usually in meters (m). Speed ​​refers to the relative speed of the obstacles, usually in meters per second (m / s). Road geometry information includes parameters such as lane curvature, road width, and slope. Curvature is usually measured in meters per second (1 / m), road width in meters (m), and slope in percentage (%). This data is obtained through the output interfaces of each sensor controller, such as via Ethernet or CAN bus, receiving the integrated environmental data processed by the perception fusion algorithm. Data reception employs periodic interrupt service routines or data callback mechanisms to ensure the real-time nature and accuracy of the environmental information. All received data undergoes coordinate transformation and data verification, uniformly converted to a vehicle coordinate system with the vehicle's center of gravity as the origin, providing a consistent environmental model for subsequent path planning. Data verification includes range checks and consistency checks, such as whether the obstacle's location coordinates are within the sensor's effective detection range, and whether different sensors produce consistent detection results for the same obstacle, ensuring the reliability and availability of environmental data.

[0024] In determining whether a drive system is in or about to enter a performance degradation state based on the consistency of frequency domain characteristics between historical load spectra and real-time operating conditions, the historical load spectra and real-time operating conditions are first transformed in the frequency domain to obtain the corresponding historical and real-time frequency domain characteristics. The frequency domain transformation is implemented using the Fast Fourier Transform (FFT) method, converting the historical load spectrum data and real-time operating condition data in the time domain into a frequency domain representation.

[0025] Historical load spectrum data comes from previously stored drive system operating records, containing sequences of torque and speed changes over time, with torque measured in Newton-meters (N·m) and speed in revolutions per minute (rpm). Real-time operating condition data includes currently acquired drive motor torque commands, actual output torque, and speed values, acquired using methods described in the preceding steps. Before frequency domain transformation, preprocessing operations are performed on the input data, including removing DC components and filtering outliers, for example, using a moving average filter with a window length of 5 sampling points to smooth data fluctuations. During the transformation, appropriate sampling frequencies and window functions are set. The sampling frequency is typically selected based on the characteristics of the drive system, for example, set to 1000 Hz. The Hanning window can be selected to reduce spectral leakage. The historical and real-time frequency domain features obtained after frequency domain transformation contain amplitude and phase spectrum information. These features represent the intensity distribution of each frequency component in complex form. The frequency resolution is determined by the sampling frequency and the number of sampling points. For example, when the sampling frequency is 1000 Hz and the number of sampling points is 1024, the frequency resolution is approximately 0.98 Hz.

[0026] Next, the similarity of the distribution of historical and real-time frequency domain features in the key frequency band is calculated as the frequency domain feature conformity. The key frequency band refers to the frequency range sensitive to the operating characteristics of the drive system, determined by analyzing the frequency bands with concentrated energy in historical data, such as the 50 Hz to 200 Hz band. This frequency band range is pre-calibrated based on the mechanical resonance and electromagnetic characteristics of the drive system. The distribution similarity calculation specifically includes extracting the power spectral density distribution of historical and real-time operating condition signals within the main energy frequency band. The power spectral density is obtained by squaring the amplitude spectrum of the frequency domain features and dividing by the frequency resolution. The correlation coefficient of the power spectral density distribution of historical and real-time operating condition signals in the same key frequency band is calculated. The correlation coefficient is calculated using the Pearson correlation coefficient method, which is obtained by calculating the ratio of the product of the covariance and the standard deviation of the two power spectral density sequences. The calculation formula is: the correlation coefficient equals the covariance divided by the product of the standard deviations of the two sequences. The calculated result ranges from -1 to +1. The calculated power spectral density distribution correlation coefficient is used as a quantitative indicator of frequency domain feature conformity. The closer the value of this indicator is to positive 1, the higher the similarity between historical and real-time frequency domain features. When the correlation coefficient is lower than the set threshold, it indicates that the system operating characteristics have changed significantly.

[0027] Finally, a status judgment is made based on the frequency domain feature compliance. If the frequency domain feature compliance is lower than a preset compliance threshold, the drive system is determined to be in or about to enter a performance degradation state. The preset compliance threshold is determined through statistical analysis, calculated based on a large amount of historical data under normal conditions. For example, 1000 sets of frequency domain feature compliance data under normal operating conditions are collected, and the 5th percentile of these data distributions is taken as the threshold. The typical value range is between 0.7 and 0.9, and the specific value is adjusted according to the drive system type and application scenario. A hysteresis interval is set during the judgment process to avoid frequent state switching. For example, when the frequency domain feature compliance is lower than the threshold of 0.8, it is determined to be in a performance degradation state, and it needs to recover to above 0.85 to be released from this state. The judgment result is output as a binary status flag, and the judgment time point and the corresponding frequency domain feature compliance value are recorded to provide a decision-making basis for subsequent thermal management strategies. The entire judgment process is executed at a fixed period, such as once every 100 milliseconds, to ensure timely detection of changes in the drive system status. The execution process also includes an exception handling mechanism. When the input data is abnormal or the calculation fails, the previous valid judgment result is retained, and a fault code is recorded.

[0028] The above steps enable the determination of drive system performance degradation status based on frequency domain characteristic conformity, providing accurate status input for subsequent thermal management strategy formulation. Frequency domain analysis methods can effectively capture changes in drive system operating characteristics, promptly identify performance degradation trends, and ensure the safe operation of electric vehicles.

[0029] When a system is in or about to enter a performance degradation state, the process of generating a thermal management strategy guide involves the following steps: First, the thermal relaxation time is estimated based on the thermal capacity and thermal resistance parameters of the drive system to characterize its thermal inertia. The thermal capacity parameter represents the amount of heat absorbed by the drive system per unit temperature rise, expressed in joules per degree Celsius (J / ℃). The thermal resistance parameter represents the temperature rise per unit heat flux, expressed in degrees Celsius per watt (℃ / W). These parameters are obtained from the drive system's design specifications or through experimental testing. Experimental testing can employ a step heating method, where a constant power is applied to the drive system, and the temperature change curve is recorded. The thermal capacity and thermal resistance parameters are obtained through curve fitting. The thermal relaxation time is calculated by multiplying the thermal capacity and thermal resistance, expressed in seconds (s). For example, when the thermal capacity is 200 J / ℃ and the thermal resistance is 0.5℃ / W, the thermal relaxation time is calculated to be 100 s. This parameter reflects the response speed of the drive system to temperature changes; a larger value indicates greater system thermal inertia.

[0030] Next, the obstacle avoidance action time window is calculated based on the relative motion relationships of dynamic obstacles in the vehicle's operating environment data to characterize the urgency of the obstacle avoidance action. The vehicle's operating environment data includes the position and speed information of surrounding obstacles, acquired through environmental perception sensors such as millimeter-wave radar and lidar. The obstacle avoidance action time window is obtained by dividing the relative distance between the vehicle and the obstacle by the relative speed, measured in seconds (s). For example, when the relative distance is 10m and the relative speed is 5m / s, the obstacle avoidance action time window is calculated to be 2s. During the calculation, multiple obstacles are considered, and the smallest obstacle avoidance action time window is selected as the final value. A minimum time window threshold, such as 0.5s, is set to avoid unreasonable situations caused by excessively small calculated values. This threshold is determined based on the vehicle's braking performance and the control system's response time.

[0031] The ratio of thermal relaxation time to obstacle avoidance action time window is then used as the time-domain trade-off factor. This calculation involves a simple division operation; for example, when the thermal relaxation time is 100s and the obstacle avoidance action time window is 2s, the time-domain trade-off factor is calculated to be 50. The time-domain trade-off factor is a dimensionless value, and its magnitude reflects the relative urgency of thermal management needs and obstacle avoidance needs. A larger value indicates a higher relative urgency of obstacle avoidance actions. A division-to-zero protection mechanism is set during the calculation process. When the obstacle avoidance action time window approaches zero, the time-domain trade-off factor is set to a maximum value, such as 1000, to ensure the stability of the calculation process.

[0032] Finally, based on the continuous numerical range of the time-domain trade-off factor, a thermal management strategy guide that smoothly transitions between thermal protection and obstacle avoidance performance is dynamically generated. The value range of the time-domain trade-off factor is divided into three continuous intervals: 0 to 30 is the thermal protection priority interval, 30 to 70 is the balance interval, and above 70 is the obstacle avoidance performance priority interval. Within each interval, a linear interpolation method is used to calculate the degree of preference of the thermal management strategy. For example, in the thermal protection priority interval, the thermal protection weight linearly decreases from 1.0 to 0.7, while the obstacle avoidance performance weight linearly increases from 0.0 to 0.3. The generated thermal management strategy guide includes target temperature limits and power output limits, which are calculated based on weighting coefficients. For example, the maximum allowable temperature decreases from 120℃ in the normal state to 100℃ in the thermal protection priority state, and the maximum output power decreases from 100kW in the normal state to 70kW in the thermal protection priority state. The entire evaluation process is performed at fixed intervals, such as updating the thermal management strategy guidance every 100ms to ensure timely response to changes in system status. At the same time, a first-order low-pass filter is used to smooth the strategy parameters to avoid parameter abrupt changes that could lead to instability in the control system.

[0033] The process of constructing an action feasibility space that simultaneously satisfies internal dynamic constraints and external spatial constraints based on thermal management strategy guidance and vehicle operating environment data, and obtaining system performance constraint evaluation results by analyzing its dynamic envelope, includes the following steps. First, the thermal management strategy guidance is mapped to constraint ranges for the vehicle's longitudinal and lateral accelerations to form internal dynamic constraints. The thermal management strategy guidance includes thermal protection weight coefficients and obstacle avoidance performance weight coefficients, which are obtained through time-domain trade-off evaluation. The specific implementation of the time-domain trade-off evaluation is described in the preceding steps. The mapping process uses a linear interpolation method. For example, when the thermal protection weight coefficient is 0.8, the longitudinal acceleration limit range is adjusted from ±3.0 m / s² in the normal state to ±1.5 m / s², and the lateral acceleration limit range is adjusted from ±2.0 m / s² in the normal state to ±1.0 m / s². The calculation of the acceleration constraint range is based on vehicle dynamic characteristics and the thermal state of the drive system, ensuring that the vehicle's dynamic performance is limited in thermal protection mode to avoid overheating. The interpolation coefficients are calibrated based on experimental data, for example, by determining the acceleration limit values ​​under different thermal states through bench testing.

[0034] Next, the geometric boundaries of the drivable area are generated based on obstacle locations, road boundaries, and traffic rules from the vehicle's operating environment data to form external spatial constraints. Obstacle locations are obtained through environmental perception sensors, including obstacle contour point cloud data detected by millimeter-wave radar and lidar. The point cloud data undergoes clustering and classification to obtain obstacle boundary information. Road boundaries are obtained by fusing lane line information acquired by visual sensors with high-precision map data. The fusion algorithm uses a Kalman filter to correlate visual detection results with map data and estimate state. Traffic rules, including speed limit signs and no-overtaking zones, are obtained through visual recognition and vehicle-to-everything (V2X) communication. Visual recognition uses deep learning algorithms to detect traffic signs, while V2X communication uses DSRC or C-V2X technology to receive traffic control information. The geometric boundaries of the drivable area are generated using a polygon convex hull algorithm, converting obstacle locations, road boundaries, and traffic rule constraints into a series of continuous polygon vertex coordinates. These coordinates are referenced to the vehicle coordinate system, with the origin at the vehicle's centroid, the X-axis pointing in the vehicle's forward direction, and the Y-axis pointing to the left side of the vehicle.

[0035] Then, a multi-constraint optimization problem is established, with internal dynamic constraints as inequality constraints and external spatial constraints as equality constraints. Internal dynamic constraints are represented as feasible regions in the vehicle's state space. For example, the longitudinal acceleration constraint is represented by the inequality: -alongmax ≤ along ≤ alongmax, where along is the longitudinal acceleration and along_max is the maximum permissible longitudinal acceleration value. External spatial constraints are represented as spatial boundary conditions that the vehicle must adhere to, such as the vehicle profile not exceeding the boundary of the drivable polygon. These constraints are represented by the equality: f(x,y)=0, where x and y are the vehicle's position coordinates, and f is a function describing the boundary of the drivable region. The objective function of the optimization problem is set to minimize the curvature change of the vehicle's trajectory. The decision variables include the vehicle's position, velocity, and acceleration. The mathematical form of the optimization problem is to find the vehicle state sequence that minimizes the objective function while satisfying all inequality and equality constraints.

[0036] A feasible solution set for the multi-constraint optimization problem is obtained through real-time iterative solving, constructing an action feasibility space that is updated in real-time with vehicle movement. Initial values ​​of the current vehicle state for the multi-constraint optimization problem are initialized in each control cycle, including vehicle position, velocity, acceleration, and heading angle. These state values ​​are obtained through the vehicle state estimation module. Numerical optimization algorithms are used to solve for the set of vehicle states that satisfy internal dynamic inequalities and external spatial equality constraints. For example, interior-point methods or sequential quadratic programming algorithms are used. The interior-point method transforms the constrained optimization problem into an unconstrained optimization problem by introducing a barrier function, while sequential quadratic programming algorithms approximate the optimal solution by iteratively solving quadratic programming subproblems. The obtained feasible solution set is used as the action feasibility space at the current moment. This space is represented as a multi-dimensional convex set in the vehicle state space, containing all possible vehicle states that satisfy the constraints. The dimension of the action feasibility space depends on the number of decision variables in the optimization problem; for example, when the decision variables are position and velocity, the action feasibility space is a four-dimensional space.

[0037] This study analyzes the coupling relationship between the curvature change of the envelope boundary and the volume contraction rate of the action feasibility space within the prediction time domain. The prediction time domain is set to a range of 3 to 5 seconds, adjusted according to vehicle speed and environmental complexity; for example, 3 seconds in urban road environments and 5 seconds in highway environments. The curvature change of the envelope boundary is obtained by calculating the rate of curvature change of the action feasibility space boundary at each point, in units of meters per second (1 / m). The curvature calculation uses differential geometry methods, obtaining the curvature value by calculating the first and second derivatives of the boundary curve. The volume contraction rate of the envelope is obtained by calculating the rate of change of the action feasibility space volume over time, in units of cubic meters per second (m³ / s). The volume calculation uses numerical integration methods, such as Monte Carlo integration or Simpson integration. The coupling relationship analysis uses correlation analysis to calculate the correlation coefficient between the rate of curvature change and the volume contraction rate. The correlation coefficient is calculated using the Pearson correlation coefficient formula, and a correlation coefficient exceeding 0.7 is considered to indicate a strong coupling relationship.

[0038] Based on the co-evolutionary pattern of envelope boundary curvature change and envelope volume contraction rate, the system performance constraint evaluation results are determined. Time-series data of envelope boundary curvature change rate and envelope volume contraction rate are monitored, with a sampling period of 100ms and a data length of 20 sampling points corresponding to a 2s time window. A time-domain correlation model between the envelope boundary curvature change rate and envelope volume contraction rate is established, using an autoregressive moving average model to describe their dynamic relationship. Model parameters are estimated using the least squares method. When a co-deterioration pattern of increased curvature change and accelerated volume contraction is detected, an evaluation result indicating intensified system performance constraints is generated. The criteria for judging the co-deterioration pattern include: a curvature change rate exceeding 0.5 m / s and a volume contraction rate exceeding 0.2 m³ / s, with both trends consistent and lasting for more than 200ms. The evaluation result output is a performance constraint level index, ranging from 0 to 1. A larger value indicates a stricter performance constraint. The performance constraint level index is calculated using a fuzzy logic system, with the curvature change rate and volume contraction rate as inputs and the performance constraint level as the output.

[0039] The entire construction and analysis process is executed with a fixed control cycle, typically 50ms to 100ms, ensuring that the action feasibility space can promptly reflect changes in the vehicle's operating environment. All calculation processes include anomaly handling mechanisms; when the optimization problem has no solution or the calculation times out, the feasible solution set from the previous cycle is used as a backup plan, and the system state is marked as requiring manual intervention. In this way, real-time monitoring and performance evaluation of the vehicle's operating status are achieved, providing accurate input for subsequent path planning and dynamic obstacle avoidance decisions.

[0040] The process of predicting the real-time maximum output capability boundary of the drive system based on the system performance constraint evaluation results includes the following steps. First, the envelope boundary curvature change and envelope volume shrinkage rate are extracted from the system performance constraint evaluation results. The system performance constraint evaluation results are obtained by analyzing the dynamic envelope of the action feasibility space, as described in the preceding steps. The unit of envelope boundary curvature change is per meter (1 / m), representing the degree of change in the shape of the action feasibility space boundary; a larger value indicates a more rapid boundary change. The unit of envelope volume shrinkage rate is cubic meters per second (m³ / s), representing the rate of reduction of the usable range of the action feasibility space; a positive value indicates space shrinkage, and a negative value indicates space expansion. The extraction process reads the stored performance evaluation data through a data interface and verifies the validity of the data. The envelope volume shrinkage rate is typically between -5 m³ / s and 5 m³ / s; data exceeding this range is considered abnormal and processed accordingly.

[0041] Next, the change in the curvature of the envelope boundary is mapped to the maximum torque limit of the drive system's sustainable output. The mapping relationship is established using a linear function combined with amplitude limiting. The form of the mapping function is: Maximum torque limit = Rated torque × (1 - k × Change in envelope curvature), where k is the mapping coefficient, which is obtained through drive system characteristic experiments.

[0042] The experimental calibration process includes testing the maximum sustainable output torque of the drive system under different curvature variations, determining the value of the mapping coefficient k through data fitting, with a typical range of 0.1 to 0.2 per meter. The unit of the maximum torque limit is Newton-meter (N·m). The thermal state and mechanical characteristics of the drive system are considered during the calculation to ensure that the output torque limit value is within the safe operating range of the drive system.

[0043] The envelope volume shrinkage rate is then mapped to the maximum permissible rate of change of rotational speed of the drive system. The mapping process employs a lookup table combined with linear interpolation to establish the correspondence between the envelope volume shrinkage rate and the maximum rate of change of rotational speed. The mapping table contains multiple data points; for example, when the envelope volume shrinkage rate is -5 m³ / s, the maximum rate of change of rotational speed is set to 1500 r / min / s; when the envelope volume shrinkage rate is 0 m³ / s, the maximum rate of change of rotational speed is set to 1000 r / min / s; and when the envelope volume shrinkage rate is 5 m³ / s, the maximum rate of change of rotational speed is set to 500 r / min / s. The parameters of the mapping table are determined through system testing. The testing method includes measuring the maximum rate of change of rotational speed that the drive system can withstand under different volume shrinkage rates, taking into account the inertial characteristics and control response characteristics of the drive system. The unit of the maximum rate of change of rotational speed is revolutions per minute per second (r / min / s), and the calculation uses linear interpolation to calculate the corresponding limit value based on the current envelope volume shrinkage rate.

[0044] Finally, by combining the maximum torque limit and the maximum speed change rate limit, the real-time maximum output capability boundary of the drive system in the prediction time domain is generated. The combined method employs a multi-constraint optimization approach, using the maximum torque limit and the maximum speed change rate limit as constraints to calculate the feasible operating region of the drive system in the prediction time domain. The prediction time domain is typically set to a range of 3 to 5 seconds, dynamically adjusted according to the vehicle's operating state; for example, a longer prediction time domain is used under high-speed conditions, and a shorter prediction time domain is used under urban conditions. The real-time maximum output capability boundary is represented as a feasible region on the torque-speed plane, the boundary of which is jointly determined by the maximum torque limit curve and the maximum speed change rate limit curve. The generation process includes sampling multiple operating points on the torque-speed plane, verifying whether each operating point simultaneously satisfies the maximum torque limit and the maximum speed change rate limit, and using the set of operating points that satisfy all constraints as the real-time maximum output capability boundary. The generated maximum output capability boundary data is transmitted to the control system via the vehicle bus for subsequent path planning and motion control. The entire prediction process is executed at a fixed period, for example, updated every 100 ms, ensuring real-time reflection of changes in the drive system's state. The calculation process includes an exception handling mechanism. When the input data is abnormal, a conservative default value is used as the output. For example, 70% of the rated torque is used as the maximum torque limit, and 800 r / min / s is used as the maximum speed change rate limit to ensure the safe operation of the system.

[0045] The process of replanning the dynamic obstacle avoidance path based on the real-time maximum output capability boundary and vehicle operating environment data, and controlling the electric vehicle to perform obstacle avoidance operations, includes the following steps. First, a speed profile conforming to the instantaneous power limit of the drive system is generated based on the real-time maximum output capability boundary. The real-time maximum output capability boundary, obtained through the aforementioned steps, is represented as a feasible working region on the torque-speed plane, defined by the maximum torque limit curve and the maximum speed change rate limit curve. The speed profile generation employs a fifth-order polynomial curve fitting method, planning a speed change curve that satisfies the drive system power limit within the prediction time domain. The prediction time domain is typically set to 3 to 5 seconds, with the specific value dynamically adjusted according to vehicle speed and environmental complexity. The speed profile generation considers the dynamic response characteristics of the drive system, ensuring that the rate of acceleration change does not exceed the system's maximum allowable value, for example, limited to within 5 m / s³. This limit is determined through drive system dynamic response testing. The generated speed profile contains time-speed sequence data, with a time interval typically set to 0.1 s, and the speed unit is meters per second (m / s). The speed profile must satisfy all constraints of the real-time maximum output capability boundary.

[0046] Next, combining obstacle location and road geometry information from the vehicle's operating environment data, an obstacle avoidance trajectory satisfying vehicle kinematics is planned under velocity profile constraints. The vehicle operating environment data includes obstacle location, velocity, and size information obtained through environmental perception sensors, as well as road boundary and lane line information. The methods for acquiring this data are described in the preceding steps. The obstacle avoidance trajectory planning employs a sampling-based method, such as the fast randomized tree algorithm, to generate a series of candidate trajectories under velocity profile constraints. The optimal trajectory is then selected based on obstacle avoidance performance and comfort metrics. The trajectory planning satisfies vehicle kinematic constraints, including maximum steering angle limits and minimum turning radius limits. For example, the maximum steering angle limit is 30 degrees, and the minimum turning radius limit is 5 meters. These parameters are determined based on the vehicle chassis characteristics. The planned obstacle avoidance trajectory contains position, velocity, and acceleration information, discretely represented at 0.1-second time intervals, and the total trajectory length matches the prediction time domain.

[0047] The obstacle avoidance trajectory is then converted into drive motor torque commands and steering system angular displacement commands. The conversion process is implemented using a vehicle inverse dynamics model, where the required drive torque is calculated based on the longitudinal acceleration information in the trajectory, and the required steering angular displacement is calculated based on the path curvature information in the trajectory. The calculation of the drive motor torque command is based on the longitudinal dynamic balance of the vehicle. The total required driving force is the sum of acceleration resistance, air resistance, rolling resistance, and slope resistance. Among them, acceleration resistance is proportional to the vehicle mass and longitudinal acceleration; air resistance is proportional to air density, air resistance coefficient, vehicle frontal area, and the square of the driving speed; rolling resistance is proportional to rolling resistance coefficient, vehicle mass, and gravitational acceleration; and slope resistance is proportional to vehicle mass, gravitational acceleration, and the sine of the road slope angle. The calculated total required driving force is combined with the effective rolling radius of the drive wheels and the overall efficiency of the transmission system to obtain the required torque command of the drive motor. The vehicle mass is obtained in real time through onboard mass sensors or state estimators; the air resistance coefficient, frontal area, and rolling resistance coefficient are obtained through wind tunnel and bench tests; the effective rolling radius of the tires is determined by vehicle design parameters; the overall efficiency of the transmission system is obtained through powertrain bench test calibration; and the road slope is obtained through inertial navigation units or high-precision map data.

[0048] The calculation of steering system angular displacement commands is based on Ackermann steering geometry. The required steering angle is calculated according to the trajectory path curvature and vehicle wheelbase. The formula is: the steering angle equals the arctangent function of the product of the vehicle wheelbase and the trajectory curvature, where the vehicle wheelbase is a fixed design parameter, and the trajectory curvature is calculated from the geometric characteristics of the planned trajectory. The converted command signal is filtered to ensure the smoothness of command changes. For example, a first-order low-pass filter is used to smooth the command sequence. The filter's cutoff frequency is set to 10Hz based on the control system's response bandwidth requirements, and the filter's time constant is calculated and determined based on the system sampling period and smoothness requirements.

[0049] Finally, corresponding commands are sent to the drive motor controller and steering system controller via the vehicle bus to execute obstacle avoidance operations. The vehicle bus uses a controller area network bus or Ethernet bus, and the communication protocol follows the corresponding automotive bus standard, such as CAN 2.0B protocol or 100BASE-T1 Ethernet protocol. Commands are sent periodically, with the sending period consistent with the control system cycle, typically 10ms. The sent command data includes drive motor torque command values ​​and steering system angular displacement command values, as well as corresponding timestamps and verification information. The verification information uses a cyclic redundancy check code to ensure the integrity of data transmission. A safety monitoring mechanism is included during command execution, monitoring the command execution status in real time. When an execution anomaly is detected, corresponding fault handling procedures are initiated. For example, an alarm is triggered when the actual torque deviates from the command torque by more than 10%, and corrective control is performed when the steering angle tracking error exceeds 2 degrees. Throughout the obstacle avoidance operation, vehicle status and environmental changes are continuously monitored, and the obstacle avoidance strategy is adjusted in a timely manner as needed to ensure the safety and effectiveness of the obstacle avoidance operation. All control commands and status information are recorded in non-volatile memory for subsequent analysis and optimization.

[0050] All calculations involved in the embodiments are dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0051] It should be noted that this invention can be deployed on the device itself to realize embedded applications, or it can run on a PC or other terminal with a user interface, thereby meeting various hardware environments and usage requirements.

[0052] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wireless or wired transmission; wired transmission methods include optical fiber, twisted pair, coaxial cable, etc.; wireless transmission includes infrared, microwave, etc. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center containing one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0053] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0054] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0055] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0056] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0057] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0058] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0059] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for autonomous driving path planning and dynamic obstacle avoidance for electric vehicles, characterized in that, include: S1. Real-time acquisition of historical load spectrum, real-time operating conditions and vehicle operating environment data of electric vehicle drive system; S2. Based on the frequency domain characteristics of the historical load spectrum and the real-time operating conditions, determine whether the drive system is in or about to enter a performance degradation state. S3. When the system is in or about to enter a state of performance degradation, a time-domain trade-off evaluation is performed based on the thermal inertia characteristics of the drive system and the urgency of obstacle avoidance actions to generate a thermal management strategy guide. S4. Based on the thermal management strategy and vehicle operating environment data, construct an action feasibility space that simultaneously satisfies internal dynamic constraints and external spatial constraints, and obtain the system performance constraint evaluation results by analyzing its dynamic envelope. S5. Predict the real-time maximum output capability boundary of the drive system based on the system performance constraint evaluation results. S6. Based on the real-time maximum output capacity boundary and vehicle operating environment data, replan the dynamic obstacle avoidance path and control the electric vehicle to perform obstacle avoidance operations.

2. The autonomous driving path planning and dynamic obstacle avoidance method for electric vehicles according to claim 1, characterized in that, Real-time acquisition of historical load spectrum, real-time operating conditions, and vehicle operating environment data of the electric vehicle drive system, including: Obtain pre-stored load spectrum data of the drive system within historical operating cycles; The torque command, actual output torque, and speed of the drive motor at the current moment are collected as real-time operating conditions. The vehicle receives information on the location, speed, and road geometry of surrounding obstacles from environmental perception sensors as data for its operating environment.

3. The autonomous driving path planning and dynamic obstacle avoidance method for electric vehicles according to claim 1, characterized in that, Based on the consistency between the historical load spectrum and the frequency domain characteristics of real-time operating conditions, determine whether the drive system is in or about to enter a performance degradation state, including: The historical load spectrum and the real-time operating conditions are transformed in the frequency domain to obtain the corresponding historical frequency domain characteristics and real-time frequency domain characteristics. The similarity of the distribution of historical frequency domain features and real-time frequency domain features in key frequency bands is calculated as the frequency domain feature conformity: the power spectral density distribution of historical and real-time operating condition signals in the main energy frequency band is extracted; the correlation coefficient of the power spectral density distribution of historical and real-time operating condition signals in the same key frequency band is calculated; and the correlation coefficient of the power spectral density distribution is used as a quantitative indicator of the frequency domain feature conformity. If the frequency domain feature compliance is lower than the preset compliance threshold, the drive system is determined to be in or about to enter a performance degradation state.

4. The autonomous driving path planning and dynamic obstacle avoidance method for electric vehicles according to claim 1, characterized in that, When a system is in or about to enter a performance degradation state, a time-domain trade-off assessment is performed based on the thermal inertia characteristics of the drive system and the urgency of obstacle avoidance actions to generate a thermal management strategy guide, including: The thermal relaxation time of the drive system is estimated based on its thermal capacity and thermal resistance parameters to characterize its thermal inertia properties. The obstacle avoidance action time window is calculated based on the relative motion relationship of dynamic obstacles in the vehicle operating environment data to characterize the urgency of the obstacle avoidance action; The ratio of thermal relaxation time to obstacle avoidance action time window is used as a time-domain trade-off factor. Based on the continuous numerical range of the time-domain trade-off factor, a thermal management strategy guide that smoothly transitions between thermal protection and obstacle avoidance performance is dynamically generated.

5. The autonomous driving path planning and dynamic obstacle avoidance method for electric vehicles according to claim 4, characterized in that, The smaller the time-domain trade-off factor, the more the thermal management strategy tends to prioritize thermal protection; the larger the time-domain trade-off factor, the more the thermal management strategy tends to prioritize obstacle avoidance performance.

6. The autonomous driving path planning and dynamic obstacle avoidance method for electric vehicles according to claim 1, characterized in that, Based on thermal management strategy guidance and vehicle operating environment data, a feasibility space for action that simultaneously satisfies internal dynamic constraints and external spatial constraints is constructed. The system performance constraint evaluation results are obtained by analyzing its dynamic envelope, including: The thermal management strategy is mapped to a range of constraints on the vehicle's longitudinal and lateral accelerations to form internal dynamic constraints; The geometric boundaries of the vehicle's drivable area are generated based on the obstacle locations, road boundaries, and traffic rules in the vehicle's operating environment data to form external spatial constraints. Establish a multi-constraint optimization problem with internal dynamic constraints as inequality constraints and external space constraints as equality constraints; By iteratively solving the feasible solution set of the multi-constraint optimization problem in real time, a feasibility space for action is constructed that is updated in real time with the movement of the vehicle. The coupling relationship between the change in the curvature of the envelope boundary and the shrinkage rate of the envelope volume in the prediction time domain of the action feasibility space was analyzed. Based on the synergistic evolution pattern of envelope boundary curvature change and envelope volume shrinkage rate, the system performance constraint evaluation results are determined.

7. The autonomous driving path planning and dynamic obstacle avoidance method for electric vehicles according to claim 6, characterized in that, By iteratively solving the feasible solution set of the multi-constraint optimization problem in real time, a real-time action feasibility space is constructed as the vehicle moves. This includes: initializing the current vehicle state initial value of the multi-constraint optimization problem in each control cycle; using numerical optimization algorithms to solve for the set of vehicle states that satisfy the internal dynamic inequality constraints and the external spatial equality constraints; and using the obtained feasible solution set as the action feasibility space at the current moment.

8. The autonomous driving path planning and dynamic obstacle avoidance method for electric vehicles according to claim 6, characterized in that, Based on the co-evolutionary pattern of envelope boundary curvature change and envelope volume shrinkage rate, the system performance constraint assessment results are determined, including: monitoring time-series data of envelope boundary curvature change rate and envelope volume shrinkage rate; establishing a time-domain correlation model between envelope boundary curvature change rate and envelope volume shrinkage rate; and generating an assessment result of intensified system performance constraints when a co-deterioration pattern of increased curvature change and accelerated volume shrinkage is detected.

9. The autonomous driving path planning and dynamic obstacle avoidance method for electric vehicles according to claim 1, characterized in that, Predict the real-time maximum output capability boundary of the drive system based on the system performance constraint evaluation results, including: Extract the changes in envelope boundary curvature and envelope volume shrinkage rate from the system performance constraint evaluation results; Map the curvature variation of the envelope boundary to the maximum torque limit for the sustainable output of the drive system; Map the envelope volume shrinkage rate to the maximum allowable rate of change of rotational speed of the drive system; By combining the maximum torque limit and the maximum speed change rate limit, the real-time maximum output capability boundary of the drive system in the prediction time domain is generated.

10. The autonomous driving path planning and dynamic obstacle avoidance method for electric vehicles according to claim 1, characterized in that, Based on real-time maximum output capacity boundaries and vehicle operating environment data, a dynamic obstacle avoidance path is replanned, and the electric vehicle is controlled to perform obstacle avoidance operations, including: A speed profile that conforms to the instantaneous power limit of the drive system is generated based on the real-time maximum output capability boundary. By combining obstacle location and road geometry information from vehicle operating environment data, an obstacle avoidance trajectory that satisfies vehicle kinematics is planned under velocity profile constraints. The obstacle avoidance trajectory is converted into drive motor torque commands and steering system angular displacement commands; The vehicle bus sends corresponding commands to the drive motor controller and steering system controller to perform obstacle avoidance operations.

Images