A road condition self-adaptive based electric wheelchair walking assisting control method and system

Abstract

The present application relates to the technical field of wheelchair control, and particularly relates to an electric wheelchair walking assisting control method and system based on road condition self-adaptation, which comprises the following steps: acquiring internal electrical signals and bilateral rotation speed feedback signals, collecting acceleration and angular velocity information of the wheelchair; performing fusion processing on the acceleration and angular velocity information to obtain a pitch angle; constructing an intention feature quantity reflecting a user's pushing intention and load change through an intention recognition algorithm to generate a reference assisting force; performing feedforward compensation processing on the reference assisting force to adapt to the type of the driving road to obtain feedforward compensation torques of bilateral motors; performing synchronization dynamic calibration on the feedforward compensation torques of the bilateral motors through a synchronization calibration algorithm; superimposing the reference assisting force and the calibrated feedforward compensation torques through a torque calculation model to obtain target torques and obtain motor control instructions. Through the fusion of motor operation parameters and IMU attitude data, the user's pushing intention is accurately recognized and the road condition is self-adapted, and the safety is significantly improved.

Description

Technical Field

[0001] This invention relates to the field of wheelchair control technology, and in particular to an electric wheelchair assistive control method and system based on road condition adaptation. Background Technology

[0002] In the field of intelligent control of electric wheelchairs, the practicality, safety and economy of the assistive mode are the core research and development directions. The current mainstream electric wheelchair assistive control solutions have obvious technical shortcomings and are difficult to meet the diverse needs of users.

[0003] Some systems directly detect the user's thrust using external torque or pressure sensors, and then drive a motor to provide assistance based on the sensor signals. While this type of solution can directly respond to thrust, it suffers from complex structure, high hardware costs, and the sensors have stringent installation accuracy requirements. They are also susceptible to aging and inaccuracy due to environmental factors such as humidity and vibration, resulting in a significant decrease in reliability and a corresponding increase in maintenance costs after long-term use.

[0004] Other systems employ speed-following logic, triggering a fixed proportion of assistance only through minute changes in the wheelchair's speed when it is pushed. This solution requires no additional sensors and is low-cost, but its core functionality has significant flaws. It cannot quantify the strength of the user's pushing intention, resulting in a lack of targeted assistance output. Furthermore, it lacks road condition awareness, failing to recognize uphill or downhill scenarios. This leads to insufficient assistance uphill, increasing the burden on the user, and a tendency for acceleration and gliding downhill, posing safety hazards. Overall, its handling comfort and safety are subpar.

[0005] Furthermore, existing solutions do not adequately consider the impact of motor parameter drift and ambient temperature changes on control accuracy, further reducing the stability of the assistive mode. Therefore, there is an urgent need for an electric wheelchair assistive control solution that does not require external force sensors, can accurately recognize user intentions, and adapts to different road conditions to address the shortcomings of existing technologies. Summary of the Invention

[0006] This invention eliminates the need for external force sensors. By fusing motor operating parameters and IMU attitude data, it accurately identifies the user's pushing intention and adapts to road conditions. This simplifies the structure, reduces costs, and significantly improves the safety, comfort, and stability of the walking aid, thus meeting diverse usage needs.

[0007] The technical solution proposed in this invention is: a road condition adaptive electric wheelchair assistive control method, the method comprising: The internal electrical signals and bilateral speed feedback signals of the motor control components are acquired, while the acceleration and angular velocity information of the wheelchair are collected by the attitude detection unit. The pitch angle is obtained by fusing acceleration and angular velocity information using an attitude fusion algorithm. Based on internal electrical signals and dual-sided speed feedback signals, an intention feature quantity reflecting the user's pushing intention and load changes is constructed through an intention recognition algorithm to generate a baseline assist. The road condition compensation model determines the type of road condition for the wheelchair based on the pitch angle, performs feedforward compensation processing on the baseline assist that is adapted to the type of road condition, and obtains the feedforward compensation torque of the dual motors. Based on the difference between the two-sided speed feedback signals and the driving road condition type corresponding to the pitch angle, the feedforward compensation torque of the two-sided motors is synchronously and dynamically calibrated using a synchronous calibration algorithm. The target torque is obtained by superimposing the reference assist torque and the calibrated feedforward compensation torque through the torque calculation model, and the target torque is converted into motor control commands through the command conversion algorithm.

[0008] Preferably, the specific process for obtaining the internal electrical signal, the dual-sided rotational speed feedback signal, the acceleration, and the angular velocity information is as follows: The operating data of the motor control components are collected and analyzed in real time to extract internal electrical signals that can characterize the motor drive state. At the same time, the speed feedback signals of the drive motors on both sides of the electric wheelchair are collected. The posture detection unit performs multi-dimensional perception of the spatial motion parameters of the electric wheelchair. Acceleration and angular velocity information of the wheelchair during movement are collected according to a preset sampling frequency. The collected motor operation parameters and posture perception parameters are synchronized to form a joint perception dataset with time dimension matching.

[0009] Preferably, the specific process for obtaining the pitch angle is as follows: The acceleration and angular velocity information in the joint sensing dataset are preprocessed, and invalid data caused by environmental interference and acquisition errors are eliminated through data filtering techniques. A multi-dimensional data fusion attitude fusion algorithm is used to perform weighted fusion and calculation on the preprocessed acceleration and angular velocity information to establish a spatial angle correlation model between the electric wheelchair body and the horizontal plane; The pitch angle, which can characterize the tilt state of the wheelchair, is calculated by a spatial angle correlation model, thus transforming multi-dimensional attitude perception data into a single road condition quantitative parameter.

[0010] Preferably, the specific process for obtaining the benchmark assistance is as follows: Feature mining is performed on the internal electrical signals and dual-side speed feedback signals in the joint sensing dataset to extract key feature parameters that can reflect changes in motor load and speed fluctuations. By fusing and modeling key feature parameters through intent recognition algorithms, an intent feature quantity that can associate motor operating status with user operation behavior is constructed, thereby realizing a quantitative representation of the strength of the user's pushing intention and the real-time load change of the wheelchair. Based on the numerical characteristics of intent features and combined with the power adaptation rules of electric wheelchairs, a quantitative calculation model for assist generation is constructed. The benchmark assist that is adapted to the user's pushing intent and actual load is calculated through the quantitative calculation model for assist generation.

[0011] Preferably, the specific process for obtaining the feedforward compensation torque is as follows: The calculated pitch angle is input into the road condition compensation model. By using the preset angle threshold judgment rules and combining the positive and negative values ​​and the range of pitch angle values, the model can accurately determine the three types of road conditions for electric wheelchairs: uphill, downhill, and flat road. To meet the power demands of different road conditions, a differentiated feedforward compensation logic is constructed, and corresponding positive compensation, negative compensation, or no compensation processing is performed on the benchmark assist. During the compensation process, the calculation rules for compensation torque are determined in combination with the characteristics of the road condition. Based on the driving coordination characteristics of the dual motors of the electric wheelchair, the calculated overall compensation torque is distributed to the two drive motors according to the preset distribution rules, and corresponding feedforward compensation torques are generated respectively.

[0012] Preferably, the specific process for obtaining the synchronous dynamic calibration of the feedforward compensation torque is as follows: The difference in speed feedback signals between the two motors on both sides of the electric wheelchair is calculated in real time to determine the degree of speed imbalance between the two motors, and the driving road condition type corresponding to the current pitch angle is retrieved at the same time. The degree of speed imbalance and the type of road conditions are used as input parameters to import into the synchronous calibration algorithm to build a torque calibration model based on road condition characteristics. This model presets differentiated calibration thresholds and fine-tuning rules according to the driving stability requirements under different road conditions. By using a torque calibration model, the feedforward compensation torque of the dual motors is dynamically fine-tuned on one side, and the compensation torque of the motor on the side with higher speed is adjusted for adaptability, thereby achieving dynamic balance of the feedforward compensation torque of the dual motors.

[0013] Preferably, the specific process for obtaining the target torque is as follows: The generated benchmark assist and the feedforward compensation torque of the dual motors after synchronous dynamic calibration are synchronously input into the torque calculation model. According to the torque fusion calculation rules preset by the model, the benchmark assist and compensation torque corresponding to the two motors are superimposed to obtain the initial target torque corresponding to each of the two motors. The initial target torque is subjected to safety limiting processing. Based on the rated operating parameters of the motor control components and the driving safety requirements of the electric wheelchair, the upper and lower threshold values ​​of the torque are set. Invalid torque values ​​that exceed the threshold range are eliminated, and the torque values ​​within the threshold range are precisely corrected to generate the target torque of both motors that meets the requirements of motor operation safety and wheelchair driving.

[0014] Preferably, the specific process of the motor control command is as follows: The target torque of the dual motors after safety limiting is input to the command conversion algorithm. The algorithm performs signal conversion and encoding processing on the numerical signal of the target torque according to the signal recognition specifications and communication protocol of the motor control component, and converts the quantized torque value into an electrical signal format that can be recognized by the motor control component. The converted electrical signal is validated to eliminate erroneous signals generated during signal transmission and encoding, ensuring the accuracy and compatibility of the control signal. The electrical signals that pass the verification are standardized and encapsulated to generate motor control commands for the two drive motors respectively, realizing the conversion from torque values ​​to executable motor control commands.

[0015] The present invention also provides a road condition adaptive electric wheelchair assistive control system, the system being used to execute the aforementioned road condition adaptive electric wheelchair assistive control method.

[0016] The present invention also provides a computer-readable storage medium storing a computer program that is executed by a processor to implement the aforementioned road condition adaptive electric wheelchair assistive control method.

[0017] The beneficial effects of this invention are: 1. This solution abandons traditional external torque and pressure sensors, and only integrates motor operating parameters (current, speed) and IMU attitude data to achieve core perception. This not only greatly simplifies the equipment structure and reduces hardware and maintenance costs, but also avoids reliability issues caused by sensor aging and inaccuracy, providing stable data support for control accuracy.

[0018] 2. By quantifying user intentions through a linear regression model and accurately identifying road conditions by calculating pitch angles using complementary filtering, differentiated control is achieved for uphill boost, downhill braking, and smooth assist on flat roads. This completely solves the shortcomings of traditional solutions, such as ambiguous intention response and insufficient road condition adaptation, and significantly improves handling comfort and driving safety.

[0019] 3. It integrates modules such as temperature compensation, online motor parameter identification, and dual-sided torque synchronous calibration to dynamically offset the effects of environmental interference and parameter drift, ensuring stable control accuracy under different operating conditions, avoiding error accumulation during long-term operation, and further enhancing the system's operational reliability and adaptability. Attached Figure Description

[0020] Figure 1 A flowchart of a road condition-adaptive electric wheelchair mobility control method; Figure 2 A flowchart illustrating the coordinate transformation between the vehicle body and the handlebars in a road condition-adaptive electric wheelchair assistive control method. Figure 3 This is a flowchart of IMU zero bias and temperature compensation for a road condition adaptive electric wheelchair assistive control method. Figure 4 A flowchart illustrating the pitch angle calculation process for a road condition-adaptive electric wheelchair assistive control method. Figure 5 A flowchart for generating a baseline assist control method for an electric wheelchair based on road condition adaptation; Figure 6 A flowchart of feedforward compensation torque distribution for a road condition adaptive electric wheelchair assistive control method; Figure 7 This is a flowchart of the synchronous calibration process for a road condition-adaptive electric wheelchair assistive control method. Detailed Implementation

[0021] The following description is intended to disclose the present invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art. The basic principles of the invention defined in the following description can be applied to other embodiments, modifications, improvements, equivalents, and other technical solutions that do not depart from the spirit and scope of the invention.

[0022] It is understood that the term "a" should be understood as "at least one" or "one or more," that is, in one embodiment, the number of an element can be one, while in another embodiment, the number of the element can be multiple, and the term "a" should not be understood as a limitation on the number.

[0023] like Figure 1 and Figure 2As shown, an electric wheelchair assistive control method based on road condition adaptation includes: acquiring internal electrical signals and bilateral speed feedback signals from the motor control unit, while simultaneously collecting acceleration and angular velocity information of the wheelchair by an attitude detection unit; fusing the acceleration and angular velocity information using an attitude fusion algorithm to obtain the pitch angle; constructing an intention feature quantity reflecting the user's pushing intention and load changes using an intention recognition algorithm based on the internal electrical signals and bilateral speed feedback signals to generate a baseline assist; determining the road condition type of the wheelchair based on the pitch angle using a road condition compensation model, performing feedforward compensation processing on the baseline assist to match the road condition type, and obtaining the feedforward compensation torque of the bilateral motors; performing synchronous dynamic calibration of the feedforward compensation torque of the bilateral motors using a synchronous calibration algorithm based on the difference between the bilateral speed feedback signals and the road condition type corresponding to the pitch angle; superimposing the baseline assist and the calibrated feedforward compensation torque using a torque calculation model to obtain the target torque, and converting the target torque into motor control commands using a command conversion algorithm.

[0024] This method eliminates the need for external torque or pressure sensors to directly detect user thrust. It leverages multi-dimensional monitoring data from the motor control unit's operating parameters and the attitude sensing information from the attitude detection unit to achieve adaptive control of the electric wheelchair. This solves the problems of complex structure, high cost, and poor reliability associated with traditional electric wheelchair assistive control systems that rely on external mechanical sensors. It also overcomes the shortcomings of speed-following-only assistive solutions, such as inability to perceive road conditions, insufficient uphill assistance, and easy slippage and loss of control on downhill slopes. It achieves accurate recognition of the user's pushing intention and adaptive assistance and braking control based on road conditions. Furthermore, the wheelchair's operating parameters can be adjusted using debugging tools, significantly improving the controllability, safety, comfort, and reliability of the electric wheelchair's assistive mode. The entire control process is completed within an electrical signal closed loop, resulting in fast control response and high operational stability. The execution cycle of each algorithm is uniformly set to 10 milliseconds, allowing for rapid response to dynamic changes in road conditions and operating status. It is applicable to various electric wheelchair products and can be adapted to different loads and motor power requirements through parameter calibration, demonstrating broad engineering application value.

[0025] Furthermore, the specific process for obtaining internal electrical signals, bilateral speed feedback signals, acceleration, and angular velocity information is as follows: Real-time acquisition and analysis of the motor control component's operating data are performed to extract internal electrical signals that characterize the motor's driving state. Simultaneously, speed feedback signals from the drive motors on both sides of the electric wheelchair are acquired. The spatial motion parameters of the electric wheelchair are perceived in multiple dimensions through the attitude detection unit. Acceleration and angular velocity information of the wheelchair during movement are collected according to a preset sampling frequency. The acquired motor operating parameters and attitude perception parameters are synchronized to form a joint perception dataset with time-dimension matching.

[0026] Among them, the motor control component is a motor controller, and the internal electrical signal is the target current signal of the motor. The q-axis current command can be selected as the specific implementation form. The q-axis current is linearly related to the output torque of the motor. The left and right motor speed feedback signals are collected by the Hall sensors built into the motor. The motor controller calculates the actual speed by counting the number of pulses per unit time. The sampling frequency of the motor operating parameters is set to 100 Hz.

[0027] The attitude detection unit is a six-axis inertial measurement unit (IMU), built into the handle controller of the electric wheelchair and rigidly connected to it. An algorithm for transforming the vehicle body and handle coordinates eliminates attitude acquisition interference caused by user arm movements. The IMU module is powered by a stable DC regulated power supply with an output voltage ripple ≤50 mV. It transmits data to the MCU via the IIC communication protocol at a baud rate of 400 kbps.

[0028] 1.1 Detailed calculation of the rotation matrix for the coordinate transformation between the vehicle body and the handle: 1.1.1 Definition of coordinate system: Establish separate coordinate systems for the wheelchair body and the handle, both using the right-handed coordinate system rule: Vehicle coordinate system: The origin is the midpoint of the rear wheel axle of the wheelchair, the X-axis is along the direction of the wheelchair's movement, the Y-axis is perpendicular to the ground and upwards, and the Z-axis is perpendicular to the XY plane and points to the left side of the wheelchair; Handle coordinate system: The origin is the IMU mounting center, the X-axis is along the direction of the handle push rod, the Y-axis is perpendicular to the handle and pointing upwards, and the Z-axis is perpendicular to the XY plane and points to the left side of the handle.

[0029] 1.1.2 Static state verification and separation of gravitational acceleration: When the wheelchair is stationary on a horizontal surface, raw triaxial acceleration data in the handle coordinate system are collected. The data collection time is 5 seconds, the sampling frequency is 100 Hz, and a total of 500 data sets are collected. The stationary state is verified by the vector magnitude. The square of each of the three axial accelerations in each data set is calculated, and the square root of the sum is taken to obtain the vector magnitude. This magnitude should be close to the local gravitational acceleration (9.80665). When the absolute value of the deviation between the average modulus of 500 data sets and the gravitational acceleration is ≤0.1. And the single deviation exceeds 0.2 Data sets with ≤5 sets are considered valid static state data.

[0030] In a stationary state, the three-axis acceleration components of the handle coordinate system are the projections of gravitational acceleration onto each axis: X-axis acceleration equals negative gravitational acceleration multiplied by the sine of the pitch angle and then by the cosine of the roll angle; Y-axis acceleration equals gravitational acceleration multiplied by the cosine of the pitch angle and then by the cosine of the roll angle; Z-axis acceleration equals gravitational acceleration multiplied by the sine of the roll angle. The pitch angle is the angle of tilt of the handle relative to the vehicle body (-90°). 90°), the roll angle is the left and right tilt angle (-90°) 90°), yaw angle is the horizontal rotation angle (0°) (360°), the yaw angle is set to 0° by default when stationary.

[0031] 1.1.3 Installation angle calculation steps: Roll angle calculation: Calculate the average value of 500 sets of Z-axis acceleration data, remove outliers and recalculate the average value. Divide the average value by the gravitational acceleration to obtain the sine value of the roll angle. Calculate the roll angle using the arcsine function with an accuracy of 0.01°. Pitch angle calculation: Calculate the average values ​​of the X-axis and Y-axis acceleration data respectively. After removing outliers, divide the negative average X-axis acceleration by the average Y-axis acceleration to obtain the tangent value of the pitch angle. Calculate the pitch angle using the arctangent function with an accuracy of 0.01°. Yaw angle calibration: Manually rotate the handle to be consistent with the forward direction of the vehicle body, collect the Z-axis angular velocity data of the gyroscope, and when the absolute value of 100 consecutive sets of angular velocity data is ≤0.1° / second, the initial value of the yaw angle is calibrated to be 0°.

[0032] 1.1.4 Construction of the rotation matrix: Based on the calculated pitch, roll, and yaw angles, a three-dimensional rotation matrix is ​​constructed from the controller coordinate system to the vehicle coordinate system. This matrix is ​​obtained by multiplying three basic rotation matrices around the X-axis (roll), Y-axis (pitch), and Z-axis (yaw) in the order of "around the X-axis → around the Y-axis → around the Z-axis," conforming to engineering conventions for coordinate system transformation and ensuring accurate transformation logic. The acceleration and angular velocity data acquired by the IMU in the controller coordinate system are multiplied with the rotation matrix to convert them to data in the vehicle coordinate system. The computation time is ≤1 millisecond, ensuring real-time data processing.

[0033] 1.2 Zero bias and temperature compensation of the inertial measurement unit: like Figure 3As shown, the IMU sampling frequency is set to 100 Hz, and zero bias calibration is automatically performed on the first startup of each day: the system places the wheelchair on a level and flat ground as the default initial condition, automatically collects raw angular velocity data for 3 seconds, removes abnormal data exceeding ±3 times the standard deviation, and takes the average value as the zero bias value of angular velocity. Subsequent angular velocity data are all subtracted from this zero bias value to achieve zero bias compensation without any additional manual operation.

[0034] A temperature compensation module is included, and the specific correction logic is as follows: Angular velocity temperature compensation: at -10℃ Within a 50℃ operating range, a calibration temperature point was set every 5℃, for a total of 13 points. At each temperature point, after holding for 30 minutes, angular velocity data was collected for 5 seconds. The zero-bias offset was calculated, and a curve corresponding to the temperature and the zero-bias offset was obtained through quadratic polynomial fitting (goodness of fit). ≥0.98); In actual use, the temperature sensor built into the IMU (measuring range -40℃) is used. The operating temperature is collected at 85℃ (accuracy ±0.5℃). The zero offset is calculated based on the fitted curve, and the original angular velocity data is subtracted from the offset to complete the correction. Acceleration temperature correction: The correction factor is equal to 1 plus 0.001 multiplied by the difference between the current temperature and 25℃. The same factor is used for the X / Y / Z axes. The corrected data is obtained by multiplying the original acceleration data by this factor.

[0035] The preset sampling frequency is 100 Hz. The synchronization process adopts a dual verification method of hardware-triggered synchronization and software timestamp: the hardware-triggered synchronization is generated by the motor control component to generate a 100 Hz synchronization signal, and both the IMU and the motor Hall sensor collect data on the rising edge of the signal; the software timestamp adds millisecond-level timestamps to the parameters. For asynchronous data with a timestamp difference of more than 2 milliseconds, linear interpolation is used to resample to ensure that the data corresponds one-to-one in the time dimension.

[0036] The walking aid start signal recognition unit uses a capacitive touch sensor, which is installed on the left and right armrests. It adopts a dual-zone synchronous trigger plus 0.5-second long press verification to prevent accidental touch. The trigger state will automatically deactivate if it lasts for more than 5 seconds.

[0037] Furthermore, the specific process for obtaining the pitch angle is as follows: the acceleration and angular velocity information in the joint sensing dataset are preprocessed, and invalid data is eliminated through data filtering technology; the attitude fusion algorithm is used to perform weighted fusion and calculation on the preprocessed information, establish a spatial angle correlation model between the electric wheelchair body and the horizontal plane, and calculate the pitch angle that represents the tilt state of the wheelchair body.

[0038] like Figure 4As shown, the data filtering technique uses a second-order low-pass filter with a cutoff frequency of 5 Hz and a filter window size of 5 sampling points. The average value is taken through a sliding window to achieve data smoothing, effectively eliminating high-frequency vibration interference and retaining the effective signal of attitude change.

[0039] The attitude fusion algorithm uses a complementary filtering algorithm. The specific calculation principle is as follows: the pitch angle of the current frame is obtained by weighted fusion of the pitch angle of the previous frame combined with the result of the integral of angular velocity and the pitch angle result calculated by acceleration. The weight coefficient of the integral of angular velocity is set to 0.98, the weight coefficient of the acceleration calculation result is set to 0.02, and the sampling period is 0.01 seconds.

[0040] 2.1 The specific process of calculating the pitch angle using acceleration: The calculation logic for pitch angle based on acceleration is as follows: First, calculate the sum of the squares of the acceleration along the Y-axis and the Z-axis in the vehicle coordinate system, and take the square root as the denominator; then, use the acceleration along the X-axis as the numerator, and use the ratio obtained by dividing the numerator by the denominator to calculate the pitch angle using the arctangent function.

[0041] In this design, the electric wheelchair is designed for low roll conditions (roll angle ≤ 5° during daily driving). At this time, the square root of the sum of the squares of the Y-axis acceleration and the Z-axis acceleration is approximately equal to the gravitational acceleration. This can be simplified to obtaining the pitch angle by using the ratio of the X-axis acceleration to the gravitational acceleration through an arcsine function, with a simplification error ≤ 0.1°. The initial condition is that the wheelchair is stationary on a horizontal surface, and the initial pitch angle is set to 0 degrees.

[0042] 2.2 The specific process of acceleration correction for integral error: The absolute value of the deviation between the pitch angle calculated by acceleration and the pitch angle obtained by integrating angular velocity is calculated in real time. The pitch angle obtained by integrating angular velocity is equal to the pitch angle of the previous frame plus the angular velocity multiplied by 0.01 seconds. When the absolute value of the deviation exceeds 0.05 radians (approximately 2.86 degrees), an integration error correction is triggered: the angular velocity integration result is adjusted towards the acceleration calculation result by 0.1 times the absolute value of the deviation, and the direction of adjustment is determined by the sign of the difference between the two. The adjusted integration result is then used for complementary filtering. If the absolute value of the deviation is ≤0.05 radians, the original calculation rules apply.

[0043] The calculated pitch angle ranges from -30°. 30°, where θ greater than 0 indicates an uphill state and θ less than 0 indicates a downhill state. This parameter is the core input basis for the road condition compensation model to determine the road condition type. The relevant calculation parameters can be adjusted and optimized through debugging tools.

[0044] Furthermore, the specific process for obtaining the benchmark assistance is as follows: feature mining is performed on the internal electrical signals and dual-side speed feedback signals in the joint sensing dataset to extract key feature parameters such as the target current change rate, the dual-side motor speed variance, and the instantaneous speed change; the key feature parameters are fused and modeled using an intent recognition algorithm to construct intent feature quantities; based on the intent feature quantities and the electric wheelchair power adaptation rules, a quantitative calculation model for assistance generation is constructed to calculate the benchmark assistance that matches the user's pushing intent and the actual load.

[0045] like Figure 5 As shown, key feature parameters are extracted using a sliding window statistical method with 10 sampling points. The mean and rate of change of the parameters within the window are calculated, and the window corresponds to a time length of 0.1 seconds, balancing the real-time performance and stability of the features.

[0046] 3.1 Normalization rules for intent recognition algorithms: The intent recognition algorithm employs a linear regression model. The intent feature is a weighted sum of three parameters: the mean rate of change of the target current within the sliding window, the variance of the speeds of both motors, and the instantaneous change in speed. All three parameters are normalized to 0 using a min-max normalization process. For interval 1, the normalization calculation logic is that the normalized value is equal to (current parameter value - minimum parameter value) divided by (maximum parameter value - minimum parameter value).

[0047] The rated operating range extreme values ​​for each parameter are: Target current change rate: minimum 0 amperes / second, maximum is the motor rated current divided by 0.1 seconds; Dual-side motor speed variance: minimum value 0 Maximum value 100 ; Instantaneous change in speed: minimum value is 0 rpm, maximum value is 10% of the motor's rated speed.

[0048] After normalization, the weighting coefficient for the mean rate of change of the target current is set to 0.5, the weighting coefficient for the variance of the speeds of the two motors is set to 0.3, and the weighting coefficient for the instantaneous change in speed is set to 0.2. The sum of these three is 1, which can be fine-tuned using the debugging tool. The final intended feature value range is 0. 1 and 0 indicate no intention to promote, while 1 indicates the greatest intention to promote.

[0049] The quantitative calculation model for the benchmark assist is as follows: ,in, To support the benchmark, The first calibration coefficient, For the target current, This is the second calibration coefficient. At the current speed, The reference speed is the initial stationary speed of the wheelchair (fixed at 0 m / s), which can be finely adjusted using the adjustment tool. The first calibration coefficient is equal to the rated torque of the motor divided by the rated current of the motor, and the second calibration coefficient is obtained by fitting the actual vehicle test under different loads and different pushing speeds. Both can be finely adjusted using the adjustment tool.

[0050] 3.2 Online Motor Parameter Identification Module: The model has an online motor parameter identification module, which uses the recursive least squares method to identify the motor back electromotive force constant, strictly following the standard recursive least squares logic. The specific steps are as follows: Simplified motor voltage balance equation: Motor terminal voltage equals motor stator resistance multiplied by phase current plus motor inductance multiplied by current change rate plus back EMF constant multiplied by motor angular velocity. Ignoring inductance voltage drop during steady-state operation, the equation simplifies to the difference between motor terminal voltage and stator resistance multiplied by phase current, which equals back EMF constant multiplied by motor angular velocity. Here, "motor terminal voltage minus stator resistance multiplied by phase current" corresponds to the observed value y(k) in the standard algorithm, and "angular velocity" corresponds to the regression vector φ(k) in the standard algorithm. Core definition of recursive least squares method: Gain matrix K(k): equals covariance matrix P(k-1) multiplied by regression vector φ(k), then divided by (1 plus regression vector φ(k) multiplied by covariance matrix P(k-1) and then multiplied by regression vector φ(k)). The new parameter estimate θ(k) is equal to the previous parameter estimate θ(k-1) plus the gain matrix K(k) multiplied by (the observed value y(k) minus the regression vector φ(k) multiplied by the previous parameter estimate θ(k-1)). The new covariance matrix P(k) is equal to (identity matrix I minus gain matrix K(k) multiplied by regression vector φ(k)) multiplied by covariance matrix P(k-1); Substitute the specific parameters: Identify only the back electromotive force constant, the covariance matrix P is a scalar (initial value 100), the regression vector φ(k) is the angular velocity, the observed value y(k) is the motor terminal voltage minus the stator resistance multiplied by the phase current, and the identity matrix I is set to 1; Stator resistance temperature compensation: The stator resistance value at the current temperature is equal to the rated resistance value at 25℃ multiplied by (1 + 0.0039 per degree Celsius × (real-time motor temperature - 25℃)). The real-time motor temperature is acquired by a motor housing temperature sensor (measurement range -40℃). 125℃, accuracy ±1℃).

[0051] This module updates the back EMF constant every 100 milliseconds and corrects the first calibration coefficient (the correction value is equal to the initial value multiplied by the factory-calibrated back EMF constant and then divided by the identified back EMF constant), adapting to parameter drift caused by motor aging and load changes.

[0052] Furthermore, the specific process of obtaining the feedforward compensation torque is as follows: The calculated pitch angle is input into the road condition compensation model, and the driving road condition type is determined by the preset angle threshold judgment rule; Differentiated feedforward compensation logic is constructed for different road conditions, and positive, negative or no compensation processing is performed on the benchmark assist; According to the dual-motor drive coordination characteristics, the overall compensation torque is distributed to the two drive motors according to the preset distribution rule to generate the corresponding feedforward compensation torque.

[0053] like Figure 6 As shown, the road condition compensation model has a preset slope threshold, with a standard value of 3 degrees. For lightweight wheelchairs (weight ≤ 20kg), this can be lowered to 2 degrees, and for heavy-duty wheelchairs (weight ≥ 40kg), it can be raised to 4 degrees. This can be calibrated using the adjustment tool. The angle threshold determination rule is as follows: a pitch angle greater than the slope threshold is considered an uphill condition; a pitch angle less than a negative slope threshold is considered a downhill condition; and an absolute value less than or equal to the slope threshold is considered a flat road condition.

[0054] The differentiated feedforward compensation logic is as follows: Uphill condition: The slope compensation torque is equal to the uphill compensation calibration coefficient multiplied by the pitch angle (in radians), ensuring consistency of physical dimensions. The uphill compensation calibration coefficient is equal to the wheelchair's curb weight multiplied by the gravitational acceleration (mg). Since the sine value is approximately equal to the radian value at small angles, the compensation torque can accurately counteract the gravitational component. Downhill condition: The slope compensation torque is equal to the negative downhill compensation calibration coefficient multiplied by the pitch angle (in radians) and then multiplied by the current speed. The negative sign indicates the braking direction. The higher the speed, the stronger the braking force. Flat road condition: compensation torque is 0.

[0055] Both uphill and downhill compensation calibration coefficients can be adjusted and optimized using debugging tools to ensure compensation accuracy under different road conditions.

[0056] The overall compensation torque is distributed according to the load ratio distribution rule of the two motors: the load ratio coefficient of the left motor is the real-time phase current of the left motor divided by the sum of the real-time phase currents of the two motors, and the load ratio coefficient of the right motor is the real-time phase current of the right motor divided by the sum of the real-time phase currents of the two motors. The overall compensation torque is multiplied by the load ratio coefficients of both sides to obtain the feedforward compensation torque of each motor, thereby achieving torque distribution for load adaptation.

[0057] 4.1 Single-sided slippage determination: Slippage Judgment Criteria: Main criteria: The speed of a single motor exceeds 1.5 times the speed of the opposite motor, and this condition persists for more than 3 frames (30 milliseconds); Auxiliary judgment: The variance of the motor speed fluctuation on this side exceeds 50 within a 3-frame time window. ; Compensation adjustment rules: After slippage is detected, the feedforward compensation torque on the slipping side will be temporarily reduced by 30% (adjusted torque = original torque × 0.7) until the slippage is resolved (the rotational speed on one side drops to 1.2 times or less of the opposite side and lasts for 5 frames), and then the original allocation rules will be restored; if slippage lasts for more than 2 seconds, a fault prompt will be triggered (the buzzer will sound continuously and the indicator light will flash red), limiting the maximum speed of the wheelchair to 0.5 m / s.

[0058] Furthermore, the specific process for obtaining the synchronous dynamic calibration of the feedforward compensation torque is as follows: the difference between the speed feedback signals of the two motors is calculated in real time to determine the degree of speed imbalance, and the road condition type corresponding to the current pitch angle is retrieved; the degree of speed imbalance and the road condition type are imported into the synchronous calibration algorithm to construct a torque calibration model based on road condition characteristics; the feedforward compensation torque of the two motors is dynamically fine-tuned on one side through this model to achieve dynamic balance of torque on both sides.

[0059] like Figure 7 As shown, the degree of speed imbalance is quantified by a relative ratio. The calculation logic is the absolute value of the difference between the speeds of the two motors divided by the maximum value of the speeds of the two motors. This ratio can eliminate the influence of the absolute value of the speed on the imbalance judgment and is the core indicator for judging whether the power output of the two motors is balanced.

[0060] The torque calibration model employs a combination of threshold segmentation adjustment and pure proportional control, without integral or derivative components. Different calibration thresholds are preset according to different road conditions: 5% for flat roads, 8% for uphill roads, and 6% for downhill roads. These thresholds have been verified and determined through real vehicle experiments.

[0061] 5.1 Formula for fine-tuning rules: The formula for calculating the adjustment amount is as follows: The adjustment amount is equal to the original compensation torque on that side multiplied by the proportional coefficient of the corresponding road condition (0.2 for flat roads, 0.3 for uphill roads, and 0.25 for downhill roads), and then multiplied by the relative ratio of the speed imbalance (in decimal form). The adjusted torque is equal to the original compensation torque minus the adjustment amount. Load determination rules: The criteria for determining the side with a large load and low speed is that the real-time phase current of the motor is greater than that of the opposite side and the real-time speed is less than that of the opposite side. Only when both conditions are met simultaneously, the reference assist on that side is adjusted to the original reference assist multiplied by (1 + comprehensive difference coefficient × 0.2). The side with a small load and high speed keeps the original reference assist unchanged. Calibration exit conditions: When the speed imbalance decreases to less than 50% of the corresponding road condition calibration threshold (flat road ≤2.5%, uphill ≤4%, downhill ≤3%), or the calibration duration reaches 500 milliseconds, stop fine-tuning and restore the original compensation torque; if the imbalance is not relieved after 500 milliseconds of continuous calibration, a fault prompt will be triggered (the buzzer beeps intermittently and the indicator light flashes yellow), limiting the maximum speed of the wheelchair to 0.3m / s.

[0062] The debugging tool can monitor and adjust the proportional and integral coefficients of the speed loops of the left and right motors, and achieve balanced power output of the two motors in conjunction with the torque calibration model.

[0063] Furthermore, the specific process for obtaining the target torque is as follows: The reference assist and the calibrated feedforward compensation torque are synchronously input into the torque calculation model, and the initial target torques of the two motors are obtained by superimposing them according to the torque fusion calculation rules; the initial target torque is subjected to safety limiting processing, the upper and lower limit thresholds of the torque are set, invalid torque values ​​are eliminated and accurately corrected, and the target torque of the two motors that meets the requirements of motor operation safety and wheelchair driving is generated.

[0064] The torque fusion calculation rule is that the reference assist and the corresponding side feedforward compensation torque are directly superimposed. That is, the initial target torque is equal to the reference assist plus the feedforward compensation torque. The target torque of the two motors is calculated independently to adapt to their respective load conditions.

[0065] The model supports differential adjustment based on dual-side benchmarks. Specifically, it calculates the absolute values ​​of the current and speed differences between the two motors in real time, calculates the proportions of the current difference to the motor's rated current and the speed difference to the motor's rated speed, multiplies the current difference proportion by 0.6 and the speed difference proportion by 0.4, and adds them together to obtain a comprehensive difference coefficient (its value is naturally 0). 1. No additional normalization required); for the side with high load and low speed, the reference assist adjustment is the original reference assist multiplied by (1 + comprehensive difference coefficient × 0.2), with a maximum adjustment range of 20%.

[0066] Safety limiting is implemented through dual limiting of current and power: the upper limit threshold for torque is set at 80% of the motor's rated torque, and is lowered to 60% of the rated torque when the battery voltage is 10% lower than the rated value; the lower limit threshold for torque is set at 70% of the motor's maximum regenerative braking torque. The threshold settings are based on the motor's rated parameters, motor temperature rise limits (maximum casing temperature ≤ 85℃), and battery discharge protection requirements, and can be calibrated and adjusted using debugging tools.

[0067] Furthermore, the specific process of the motor control command is as follows: The target torque of both motors after safety limiting is input to the command conversion algorithm. According to the signal recognition specifications and communication protocols of the motor control components, the torque value signal is converted into an electrical signal format that the motor can recognize. The validity of the converted electrical signal is verified, and erroneous signals are eliminated. The verified electrical signals are standardized and encapsulated to generate control commands corresponding to the two drive motors, realizing the conversion from torque value to motor executable commands.

[0068] The instruction conversion algorithm uses the CAN bus communication protocol (baud rate 250kbps). The generation and conversion rules of the torque-current mapping table are as follows: theoretical conversion is performed based on the motor torque constant (target current equals target torque divided by motor torque constant), at 0... Within the rated torque range of the motor, a discrete mapping table is constructed by collecting a set of actual current values ​​every 0.1 Nm. During actual conversion, if the target torque is between two sampling points, linear interpolation is used to calculate the current value; if it exceeds the calibration range, it is treated as a boundary value. The target current value is quantized into a 16-bit unsigned integer (quantization principle is to multiply the current value by 100 and then round down), with a resolution of 0.01 amperes and a value range of 0. 65535.

[0069] 7.1 CAN bus message format: Frame type: Data frame; Frame ID: 0x180; Data field: 8 bytes, bytes 0-1 are the left motor current code (byte 0 is the high byte), bytes 2-3 are the right motor current code (byte 2 is the high byte), byte 4 is the status code (single byte), byte 5 is reserved (default 0x00), bytes 6-7 are CRC16 check value; Status code definition: 0x00 is a normal instruction, 0x01 is a calibration instruction, and 0x02 is a fault instruction.

[0070] The validity check uses a range check plus CRC16 cyclic redundancy check: First, it is determined whether the current code value is within the range of 0 to the code value corresponding to the rated current of the motor. If it exceeds the range, it is determined to be an invalid signal. Then, CRC16 check is performed on the signals within the valid range (the check range is the first 6 bytes of the data field) to remove erroneous signals.

[0071] 7.2 Communication Link and Command Execution Flow: The CAN bus communication link is constructed using twisted-pair shielded cable, with the shielding layer grounded (grounding resistance ≤ 4 ohms). The physical layer uses a TJA1050 transceiver. The command receiving module adopts an interrupt-driven reception method. An interrupt is triggered when a valid message is received on the CAN bus, with an interrupt response time ≤ 50 microseconds. The specific execution flow is as follows: Message parsing: Extract 8 bytes of data field content; Verification: Check the CRC16 check value, parse the status code and current code value, and perform the corresponding operation according to the status code; Encoding Reverse Decoding: The actual current value equals the encoded value divided by 100; Command execution: The actual current value is used as the target current input current loop PI control algorithm (proportional coefficient 0.8, integral coefficient 0.1, control cycle 10 milliseconds). The proportional coefficient and integral coefficient of this algorithm can be fine-tuned by the debugging tool to drive the motor to output the corresponding torque. Feedback Reporting: Collect parameters such as actual motor current, speed, and temperature, encapsulate them into status feedback messages according to the message format of frame ID=0x181, and upload them to the main controller.

[0072] 7.3 Instruction Fault Tolerance and Degradation Handling: Command timeout fault tolerance: The main controller sends a control command every 100 milliseconds. If the motor controller does not receive a valid command for 3 consecutive cycles (300 milliseconds), the degradation mode is activated (the target current of both motors drops to 0.5 amps, the wheelchair moves slowly at a low speed of ≤0.2 m / s, the buzzer sounds every 1 second, and the indicator light stays on yellow). Encoding error handling: When the current encoding value exceeds the rated range, it is automatically corrected to 80% of the rated current, marked as "over-limit correction" and uploaded, and the main controller records the abnormal log. Communication failure recovery: After the instruction transmission is restored to normal (verification passes for 5 consecutive cycles), it linearly transitions to normal control logic within 1 second.

[0073] The processes described above with reference to the flowcharts in the embodiments disclosed in this invention can be implemented as computer software programs. The embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication component, and / or installed from a removable medium. When the computer program is executed by a central processing unit (CPU), it performs the functions defined in the methods of this application. It should be noted that the computer-readable medium described above in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wire segments, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to: wireless segments, wire segments, optical fibers, RF, etc., or any suitable combination thereof.

[0074] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0075] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are merely examples and do not limit the present invention. The purpose of the present invention has been fully and effectively achieved. The functions and structural principles of the present invention have been shown and explained in the embodiments. Without departing from the stated principles, the implementation of the present invention may have any variations or modifications.

Claims

1. A road condition-based adaptive electric wheelchair walking assistance control method, characterized by, The method includes: The internal electrical signals and bilateral speed feedback signals of the motor control components are acquired, while the acceleration and angular velocity information of the wheelchair are collected by the attitude detection unit. The pitch angle is obtained by fusing acceleration and angular velocity information using an attitude fusion algorithm. Based on internal electrical signals and dual-sided speed feedback signals, an intention feature quantity reflecting the user's pushing intention and load changes is constructed through an intention recognition algorithm to generate a baseline assist. The road condition compensation model determines the type of road condition for the wheelchair based on the pitch angle, performs feedforward compensation processing on the baseline assist that is adapted to the type of road condition, and obtains the feedforward compensation torque of the dual motors. Based on the difference between the two-sided speed feedback signals and the driving road condition type corresponding to the pitch angle, the feedforward compensation torque of the two-sided motors is synchronously and dynamically calibrated using a synchronous calibration algorithm. The target torque is obtained by superimposing the reference assist torque and the calibrated feedforward compensation torque through the torque calculation model, and the target torque is converted into motor control commands through the command conversion algorithm.

2. The road condition adaptive electric wheelchair walking assistance control method according to claim 1, characterized in that, The specific process for obtaining the internal electrical signal, the dual-sided rotational speed feedback signal, the acceleration, and the angular velocity information is as follows: The operating data of the motor control components are collected and analyzed in real time to extract internal electrical signals that can characterize the motor drive state. At the same time, the speed feedback signals of the drive motors on both sides of the electric wheelchair are collected. The posture detection unit performs multi-dimensional perception of the spatial motion parameters of the electric wheelchair. Acceleration and angular velocity information of the wheelchair during movement are collected according to a preset sampling frequency. The collected motor operation parameters and posture perception parameters are synchronized to form a joint perception dataset with time dimension matching.

3. The road condition adaptive electric wheelchair walking assistance control method according to claim 2, characterized in that, The specific process for obtaining the pitch angle is as follows: The acceleration and angular velocity information in the joint sensing dataset are preprocessed, and invalid data caused by environmental interference and acquisition errors are eliminated through data filtering techniques. A multi-dimensional data fusion attitude fusion algorithm is used to perform weighted fusion and calculation on the preprocessed acceleration and angular velocity information to establish a spatial angle correlation model between the electric wheelchair body and the horizontal plane; The pitch angle, which can characterize the tilt state of the wheelchair, is calculated by a spatial angle correlation model, thus transforming multi-dimensional attitude perception data into a single road condition quantitative parameter.

4. The road condition adaptive electric wheelchair walking assistance control method according to claim 3, characterized in that, The specific process for obtaining the benchmark assist is as follows: Feature mining is performed on the internal electrical signals and dual-side speed feedback signals in the joint sensing dataset to extract key feature parameters that can reflect changes in motor load and speed fluctuations. By fusing and modeling key feature parameters through intent recognition algorithms, an intent feature quantity that can associate motor operating status with user operation behavior is constructed, thereby realizing a quantitative representation of the strength of the user's pushing intention and the real-time load change of the wheelchair. Based on the numerical characteristics of intent features and combined with the power adaptation rules of electric wheelchairs, a quantitative calculation model for assist generation is constructed. The benchmark assist that is adapted to the user's pushing intent and actual load is calculated through the quantitative calculation model for assist generation.

5. The road condition adaptive electric wheelchair walking assistance control method according to claim 4, characterized in that, The specific process for obtaining the feedforward compensation torque is as follows: The calculated pitch angle is input into the road condition compensation model. By using the preset angle threshold judgment rules and combining the positive and negative values ​​and the range of pitch angle values, the model can accurately determine the three types of road conditions for electric wheelchairs: uphill, downhill, and flat road. To meet the power demands of different road conditions, a differentiated feedforward compensation logic is constructed, and corresponding positive compensation, negative compensation, or no compensation processing is performed on the benchmark assist. During the compensation process, the calculation rules for compensation torque are determined in combination with the characteristics of the road condition. Based on the driving coordination characteristics of the dual motors of the electric wheelchair, the calculated overall compensation torque is distributed to the two drive motors according to the preset distribution rules, and corresponding feedforward compensation torques are generated respectively.

6. The road condition adaptive electric wheelchair walking assistance control method according to claim 5, wherein, The specific process for obtaining the synchronous dynamic calibration of the feedforward compensation torque is as follows: The difference in speed feedback signals between the two motors on both sides of the electric wheelchair is calculated in real time to determine the degree of speed imbalance between the two motors, and the driving road condition type corresponding to the current pitch angle is retrieved at the same time. The degree of speed imbalance and the type of road conditions are used as input parameters to import into the synchronous calibration algorithm to build a torque calibration model based on road condition characteristics. This model presets differentiated calibration thresholds and fine-tuning rules according to the driving stability requirements under different road conditions. By using a torque calibration model, the feedforward compensation torque of the dual motors is dynamically fine-tuned on one side, and the compensation torque of the motor on the side with higher speed is adjusted for adaptability, thereby achieving dynamic balance of the feedforward compensation torque of the dual motors.

7. The road condition adaptive electric wheelchair walking assistance control method according to claim 6, characterized in that, The specific process for obtaining the target torque is as follows: The generated benchmark assist and the feedforward compensation torque of the dual motors after synchronous dynamic calibration are synchronously input into the torque calculation model. According to the torque fusion calculation rules preset by the model, the benchmark assist and compensation torque corresponding to the two motors are superimposed to obtain the initial target torque corresponding to each of the two motors. The initial target torque is subjected to safety limiting processing. Based on the rated operating parameters of the motor control components and the driving safety requirements of the electric wheelchair, the upper and lower threshold values ​​of the torque are set. Invalid torque values ​​that exceed the threshold range are eliminated, and the torque values ​​within the threshold range are precisely corrected to generate the target torque of both motors that meets the requirements of motor operation safety and wheelchair driving.

8. The road condition adaptive electric wheelchair walking assistance control method according to claim 7, characterized in that, The specific process of the motor control command is as follows: The target torque of the dual motors after safety limiting is input to the command conversion algorithm. The algorithm performs signal conversion and encoding processing on the numerical signal of the target torque according to the signal recognition specifications and communication protocol of the motor control component, and converts the quantized torque value into an electrical signal format that can be recognized by the motor control component. The converted electrical signal is validated to eliminate erroneous signals generated during signal transmission and encoding, ensuring the accuracy and compatibility of the control signal. The electrical signals that pass the verification are standardized and encapsulated to generate motor control commands for the two drive motors respectively, realizing the conversion from torque values ​​to executable motor control commands.

9. A road condition adaptive electric wheelchair assistive control system, characterized in that, The system is used to execute the road condition adaptive electric wheelchair mobility control method according to any one of claims 1-8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that is executed by a processor to implement the road condition adaptive electric wheelchair assistive control method according to any one of claims 1-8.

Images