A non-inductive resistance lightweight cycling bidirectional kinetic energy recovery self-help bicycle system

Abstract

The application discloses a kind of inductive resistance lightweight cycling bidirectional energy recovery self-help force bicycle systems, comprising: power-assisted wheel hub motor integrated one-way overrunning clutch, mechanical cut-off power when not powered;MCU operation controller is embedded multi-modal finite state machine, weighted fusion filter and braking torque coordination algorithm, receive real-time signal, energy storage battery package is built-in battery management system containing temperature monitoring and overcurrent protection, by bidirectional DC converter and motor form charge-discharge circuit, controller will signal generation cycling difficulty index, drive machine in pure human power idle, help to establish, steady-state assistance, recovery establishment, steady-state recovery and braking intervention six-state switching, assist state output auxiliary torque, recovery state without brake executes perturbation and observation method maximum power point tracking, brake signal is received then with highest priority jump brake intervention state, braking torque coordination algorithm maps regenerative braking torque, while cutting off power output and shielding maximum power point tracking.

Description

Technical Field

[0001] This invention relates to the field of low-carbon travel technology, and in particular to a lightweight, non-restrictive, bidirectional kinetic energy recovery self-powered bicycle system. Background Technology

[0003] Most existing technologies use fixed rectification or constant voltage charging for energy recovery, which cannot dynamically adjust electrical parameters according to the real-time speed and load of the hub motor. This results in the generator not being able to operate at its maximum power output point at the current speed, leading to low efficiency in converting mechanical energy into electrical energy and significant energy waste during downhill / gliding phases. Therefore, a lightweight, non-impact, bidirectional kinetic energy recovery self-powered bicycle system is proposed. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a lightweight, non-restrictive, bidirectional kinetic energy recovery self-powered bicycle system.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A lightweight, drag-free cycling system with bidirectional kinetic energy recovery and self-powered operation includes a bicycle, said bicycle comprising: Power-assisted hub motor: It integrates a one-way overrunning clutch, which is located between the motor rotor and the wheel drive unit. It is used to mechanically cut off the power transmission path when the power-assisted hub motor is not powered on, so as to achieve a riding effect without feeling resistance when riding by pure human power. MCU computing controller: It has an embedded multimodal finite state machine, weighted fusion filter and braking torque coordination algorithm module, and is electrically connected to the power steering hub motor. At the same time, it receives real-time acquisition signals from the cadence sensor, slope angle sensor, vehicle speed Hall sensor and dual-channel redundant brake lever braking sensor. Energy storage battery pack: It integrates a battery management system, which includes a battery temperature monitoring module and a charge / discharge overcurrent protection module. The energy storage battery pack forms a charge / discharge closed loop with the power hub motor through a bidirectional DC-DC converter with perturbation observation method maximum power point tracking control function. The MCU computing controller generates a cycling difficulty index by passing the collected slope angle signal, cadence signal and vehicle speed signal through the weighted fusion filter, and drives the multimodal finite state machine to switch between pure human idle state, assist establishment state, steady-state assist state, recovery establishment state, steady-state recovery state and braking intervention state with anti-shake delay based on the cycling difficulty index and Schmidt hysteresis comparator. When in the boost-on state or steady-state boost state, the MCU computing controller outputs auxiliary torque to the boost wheel hub motor; When the bidirectional DC-DC converter is in the recovery establishment state or steady-state recovery state and no braking signal is received, it performs maximum power point tracking using the perturbation-observation method. When the brake signal, which has been confirmed by de-jittering, is received from the dual-redundant brake lever brake sensor, the multimodal finite state machine is forced to jump to the brake intervention state with the highest priority. The braking torque coordination algorithm module regenerates the braking torque by linear mapping based on the real-time vehicle speed signal and the brake stroke signal, and at the same time cuts off the power assist output of the power assist hub motor and shields the maximum power point tracking.

[0006] The above technical solution further includes: Furthermore, the MCU computing controller generates a cycling difficulty index through the weighted fusion filter. The weighted fusion filter is a Kalman filter, which dynamically weights and fuses the real-time slope angle signal output by the slope angle sensor, the pedal frequency signal output by the cadence sensor, and the real-time vehicle speed signal output by the vehicle speed Hall sensor to output a continuous cycling difficulty index. The Schmidt hysteresis comparator sets state transition thresholds for the riding difficulty index. The state transition thresholds include a first threshold for entering the assist mode, a second threshold for exiting the assist mode, a third threshold for entering the recovery mode, and a fourth threshold for exiting the recovery mode. The first threshold for entering the assist mode is higher than the second threshold for exiting the assist mode, and the third threshold for entering the recovery mode is lower than the fourth threshold for exiting the recovery mode. All state transitions must satisfy a corresponding duration greater than the preset anti-shake delay.

[0007] Furthermore, the braking torque coordination algorithm module performs the following operations during the braking intervention state: The system receives two Hall signals from the dual-redundant brake lever brake sensor and only confirms a real braking signal when both signals are valid at the same time and the difference between the two signals is less than a preset error threshold. The wheel slip ratio is calculated based on the real-time vehicle speed measured by the vehicle speed Hall sensor. When the wheel slip ratio is less than the preset optimal slip ratio threshold, the regenerative braking torque is allowed to increase linearly with the braking stroke. When the slip ratio is greater than the preset optimal slip ratio threshold, the regenerative braking torque is immediately set to zero and mechanical braking is used completely. The upper limit of the regenerative braking torque decreases linearly as the vehicle speed decreases. When the vehicle speed is lower than the preset low-speed exit threshold, the regenerative braking is completely disengaged. Furthermore, under downhill braking conditions, the maximum limit of the regenerative braking torque is linearly reduced based on the negative slope value of the slope angle sensor.

[0008] Furthermore, the MCU operation controller outputs an auxiliary torque smoothed by a first-order low-pass filter to the power-assisted hub motor based on a pre-stored two-dimensional torque curve mapping table and a rider fatigue factor multiplicative correction coefficient. The pre-stored two-dimensional torque curve mapping table uses the slope angle signal and cadence signal as two-dimensional input variables and the basic auxiliary torque as the output variable. The pre-stored two-dimensional torque curve mapping table is stored in the non-volatile memory inside the MCU operation controller and contains torque values ​​of 25 grid points, including no less than 5 slope angle discrete points and 5 cadence discrete points. The real-time torque is calculated by bilinear interpolation between adjacent grid points. The cyclist fatigue factor multiplicative correction coefficient is obtained by normalizing the fatigue signal collected by the cadence sensor. Its default value is 1.0, and its range is limited to between 0.8 and 1.5. Furthermore, the rate of change of this correction coefficient is limited by a first-order low-pass filter.

[0009] Furthermore, the bidirectional DC-DC converter activates perturbation-observation maximum power point tracking only when the multimodal finite state machine is in the recovery establishment state or steady-state recovery state and has not received any braking signal; After activation, the bidirectional DC-DC converter applies a small perturbation to the pulse width modulation duty cycle every 50 milliseconds and observes the change in the input power output of the booster hub motor as a generator. If the power increases, it continues to perturb in the same direction; if the power decreases, it perturbs in the opposite direction. After multiple iterations of three to five times, it converges to the maximum power point at the current speed. Once the braking signal is confirmed, the maximum power point tracking using the disturbance observation method is immediately shielded by the hardware-level signal, and the bidirectional DC-DC converter switches to the regenerative torque command mode directly given by the braking torque coordination algorithm module.

[0010] Furthermore, the multimodal finite state machine is also provided with exit transition states, which include assisted exit transition states, recovery exit transition states, and braking exit transition states; When the conditions for exiting from the steady-state assisted state to the pure manual idle state are met, the multimodal finite state machine first jumps to the assisted exit transition state. In this state, the auxiliary torque output by the assisted hub motor linearly decays to 0 within 0.5 seconds before the one-way overrunning clutch is disengaged. When the braking signal disappears, the multimodal finite state machine first jumps from the braking intervention state to the braking exit transition state, linearly decays the regenerative braking torque to 0 within 0.5 seconds, and then decides to enter the recovery establishment state, the assist establishment state, or directly enter the pure human idle state based on the current riding difficulty index and the hysteresis comparator result.

[0011] Furthermore, the dual-redundant brake lever brake sensors are respectively installed on the left and right brake levers of the bicycle. Each brake lever is equipped with two independent Hall effect sensors from different sources, and the two signals are respectively input to the MCU computing controller through different analog-to-digital conversion channels.

[0012] Furthermore, the MCU computing controller is equipped with a hardware de-jitter filter and a software de-jitter algorithm. It continuously samples 10 times, and only if 8 or more of them are valid levels is it confirmed as a real braking signal. If only one sensor outputs a valid level and the duration is longer than the preset fault confirmation time, it is determined to be a sensor fault. At this time, the MCU computing controller illuminates the fault indicator light on the handlebar control panel without entering the braking intervention state. The handlebar control panel is located above the handlebars.

[0013] The present invention has the following beneficial effects: In this invention, when going downhill or coasting and without receiving a braking signal, the bidirectional DC-DC converter performs maximum power point tracking (MPPT) using the disturbance observation method, enabling the hub motor to operate as a generator at its maximum power output point at the current speed. This significantly improves the efficiency of converting mechanical energy into electrical energy during downhill / coasting, thereby extending the driving range with the same battery capacity or reducing the frequency of external charging. At the same time, the MPPT function is only activated under safe operating conditions and is immediately shielded after a braking signal is received, achieving a balance between energy recovery efficiency and driving safety. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the overall structure of a lightweight, non-restrictive, bidirectional kinetic energy recovery self-powered bicycle system proposed in this invention. Figure 2 This is a system block diagram of a lightweight, non-restrictive, bidirectional kinetic energy recovery self-powered bicycle system proposed in this invention. Figure 3 This is a three-dimensional structural diagram of a lightweight, non-restrictive, bidirectional kinetic energy recovery self-powered bicycle system proposed in this invention. Figure 4 for Figure 3 Enlarged diagram of point A in the diagram; Figure 5 for Figure 3 Enlarged diagram of point B in the image.

[0015] In the diagram: 1. Bicycle; 2. Power-assisted hub motor; 3. One-way overrunning clutch; 4. MCU (Microcontroller Unit); 5. Cadence sensor; 6. Inclination angle sensor; 7. Vehicle speed Hall sensor; 8. Dual-channel redundant brake lever brake sensor; 9. Energy storage battery pack; 10. Brake lever; 11. Handlebar control panel. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Please see Figures 1-5 As shown, the present invention is a lightweight, non-restrictive, bidirectional kinetic energy recovery self-powered bicycle system, comprising a bicycle 1, wherein the bicycle 1 includes: Lightweight, contactless clutch-assisted hub motor 2: It integrates a one-way overrunning clutch 3, which is located between the motor rotor and the wheel drive unit. It is used to mechanically cut off the power transmission path when the hub motor 2 is not powered on, so as to achieve a contactless riding effect when riding with pure human power. The intelligent electronic control MCU computing controller 4 is embedded with a multimodal finite state machine, a weighted fusion filter and a braking torque coordination algorithm module, and is electrically connected to the power assist hub motor 2. At the same time, it receives real-time acquisition signals from the pedal frequency sensor 5, the slope angle sensor 6, the vehicle speed Hall sensor 7 and the dual-channel redundant brake lever brake sensor 8. Lightweight energy storage battery pack 9: It integrates a battery management system, which includes a battery temperature monitoring module and a charge / discharge overcurrent protection module. The energy storage battery pack 9 forms a charge / discharge closed loop with the power hub motor 2 through a bidirectional DC-DC converter with perturbation observation method maximum power point tracking control function. The lightweight energy storage battery pack 9 is connected to the battery management system (BMS) integrated inside the MCU computing controller 4 via CAN bus or UART serial port communication, and reports the remaining battery power, single cell voltage, charging and discharging current and battery temperature data in real time. The battery temperature monitoring module in the battery management system (BMS) uses an NTC thermistor array attached to the surface of the battery cell. When the temperature at any monitoring point exceeds the preset first temperature threshold, the MCU linearly reduces the charging current of kinetic energy recovery or the output power of the motor assist. When the temperature at any monitoring point exceeds the preset second temperature threshold and the second temperature threshold is higher than the first temperature threshold, the BMS forcibly disconnects the main circuit relay of the bidirectional DC-DC charge and discharge controller and independently blocks the hub motor drive signal. In addition, when the battery temperature is lower than the preset low temperature charging prohibition threshold, the MCU will forcibly prohibit kinetic energy recovery charging regardless of whether a downhill or coasting signal is received. The MCU computing controller 4 generates a cycling difficulty index by passing the collected slope angle signal, pedal frequency signal and vehicle speed signal through the weighted fusion filter, and drives the multimodal finite state machine to switch between pure human idle state, assist establishment state, steady-state assist state, recovery establishment state, steady-state recovery state and braking intervention state with anti-shake delay based on the cycling difficulty index and Schmidt hysteresis comparator. When in the power-on state or steady-state power-on state, the MCU computing controller 4 outputs the auxiliary torque smoothed by a first-order low-pass filter to the power-on hub motor 2 according to the pre-stored two-dimensional torque curve mapping table and the rider fatigue factor multiplicative correction coefficient. When the bidirectional DC-DC converter is in the recovery establishment state or steady-state recovery state and no braking signal is received, it performs maximum power point tracking using the perturbation-observation method. When the brake signal that has been de-jittered and confirmed is received from the dual-redundant brake lever brake sensor 8, the multimodal finite state machine is forced to jump to the brake intervention state with the highest priority. The braking torque coordination algorithm module regenerates the braking torque by linear mapping according to the real-time vehicle speed signal and the brake stroke signal, and at the same time cuts off the power assist output of the power assist hub motor 2 and shields the maximum power point tracking. The bicycle 1 has a slope angle sensor 6 installed at the bottom bracket of the frame body, a speed Hall sensor 7 fixedly installed at the rear chainstay of the frame body, a power assist hub motor 2 and its matching front wheel installed at the front of the frame body, a pedal crank and a cadence sensor 5 installed in the middle of the frame body, a battery pack 9 hidden inside the seat tube of the frame body, and a smart electronic control MCU controller 4 fixedly installed therein, dual-redundant brake lever brake sensors 8 installed on the left and right brake levers respectively, a waterproof touch handlebar control panel 11 installed at the handlebar position, which includes a mode switching touch button, a battery level display area and a brake safety status indicator, and the entire external wiring area of ​​the frame body is covered with a waterproof integrated wiring harness.

[0018] The working principle of a lightweight, non-restrictive, bidirectional kinetic energy recovery self-powered bicycle system is as follows: Real-time acquisition and jitter filtering of multi-source sensor signals: The MCU operation controller simultaneously receives real-time acquisition signals from the following four types of sensors at a fixed sampling period (e.g., 10 milliseconds): cadence sensor signal (Unit: rpm); Slope angle sensor signal (Unit: degrees, uphill is positive, downhill is negative); Vehicle speed Hall sensor signal v (unit: m / s); Dual-path redundant brake lever brake sensor signal (Two redundant outputs, each outputting 0 or 1); Perform jitter filtering on each signal: sample 10 times consecutively, and if at least 8 of them are valid, they are considered valid signals; otherwise, discard the sampled values. In one embodiment, the signal acquired during a certain control cycle is: After filtering, the slope angle θ = 8°; After filtering, the pedal frequency f = 55 rpm; After filtering, the vehicle speed v = 4.0 m / s (desired vehicle speed 5.5 m / s); Braking signal: =0.2V, =0.3V, none of them were triggered; =0.1V, B_R2=0.2V, neither triggered, thus there was no braking; Weighted fusion to generate a cycling difficulty index: incorporating slope angle signals cadence signal The vehicle speed signal v is input to the weighted fusion filter. This fusion filter uses a Kalman filter or a complementary filter to output a continuous riding difficulty index R. The calculation formula when using a complementary filter is: ,in, For the preset weighting coefficients, and , This is the rider's maximum theoretical cadence (e.g., 120 rpm). The rider's desired speed (which can be obtained from historical average speed or handlebar settings); In one embodiment, R = 0.6 × 8 + 0.3 × (1 - 55 / 120) + 0.1 × (1 - 4.0 / 5.5) = 4.8 + 0.3×(1 - 0.4583) + 0.1×(1 - 0.7273) =4.9898; The weighted fusion filter can also adopt the standard recursive form of a Kalman filter, calculating the optimal estimate through state prediction and measurement update iterations. This will not be elaborated upon here; Schmidt Hysteresis Comparator and State Machine Drive: The cycling difficulty index R is input to the Schmidt hysteresis comparator, which is set with the following four preset thresholds: High threshold for entering boost mode =0.65; Low threshold for exiting assist mode =0.45 ( ); Low threshold for entering recycling mode =0.20; High threshold for exiting recycling mode =0.35 ( ); The hysteresis comparison logic is as follows: ; like If the state differs from the current state machine state, a debouncing delay judgment is performed first: this must be within three consecutive sampling periods (e.g., 30 milliseconds). State transitions are only allowed if the state remains unchanged; otherwise, the original state is maintained. The multimodal finite state machine is based on The braking signal drives the state transition, and the transition rules are as follows: Pure manual idle state: no assistance, no retraction, one-way overrunning clutch remains disengaged; Assist-on state: Entered when an assist request is received and the current state is not assist-on. In this state, the assist torque increases linearly from 0 to the target value (time constant 0.5 seconds). Steady-state assist state: Assist is established and continuously outputs auxiliary torque; Recovery Setup: Entered when a recovery request is received but no braking signal is received, the recovery power increases linearly from 0 to the MPPT tracking value (time constant 0.3 seconds). Steady-state recycling: Recycling is established, and MPPT continues to be executed; Braking intervention state: highest priority. Once a braking signal that has been confirmed by de-jittering is received, the system will immediately switch to this state, regardless of its current state. In one embodiment, R≈4.99> =0.65, the hysteresis comparator outputs a help request. The state remains unchanged for 3 consecutive sampling periods (30ms), and the state machine transitions from a purely manual idle state to a help-established state; Assist torque calculation in assisted state: When the state machine is in the assisted establishment state or steady-state assisted state, the MCU operation controller performs the following torque calculation: First, based on the current slope angle and cadence The basic auxiliary torque is obtained by referring to the pre-stored two-dimensional torque curve mapping table. The mapping table contains 5×5 grid points, with bilinear interpolation used between adjacent points: ,in , , to The torque values ​​are for four adjacent grid points; Pre-stored two-dimensional torque curve mapping table: In one embodiment, the slope angle is 8°, located between grid points 5° and 10°, and the interpolation coefficient u = (8-5) / (10-5) = 0.6; With a pedal frequency of 55 rpm, located between grid points 50 and 70, the interpolation coefficient v = (55-50) / (70-50) = 0.25; =8, =16, =4, =10; ; Secondly, calculate the multiplicative correction factor for the fatigue factor. : Or, if a cadence pressure sensor is used, then ,in This represents the real-time pedaling pressure amplitude. The baseline pedaling pressure for cyclists; Assuming the cyclist's current heart rate (HR) is 145 bpm, the maximum heart rate... =190, benchmark coefficient =1.2 (conservative cyclist) min(1.5,max(0.8,145 / 190×1.2))=0.9158; Final auxiliary torque target value: The target value is then smoothed through a first-order low-pass filter to eliminate sudden torque changes. The time constant Second, For the control period (e.g., 10 milliseconds). Smoothed The hub motor is driven by a pulse width modulation (PWM) signal to output auxiliary torque; =11.5×0.9158≈10.53; First-order low-pass filter smooths the output (assuming the torque in the previous cycle is 0, and the time step is...). =0.01s, =0.2s): After about 0.5 seconds (50 cycles). When the value approaches 10.53 N·m, the state machine transitions to a steady-state boost state. Maximum Power Point Tracking in Non-Braking Regeneration State: When the state machine is in the regeneration establishment state or steady-state regeneration state and no braking signal is received, the bidirectional DC-DC converter performs perturbation-observation maximum power point tracking (MPPT). The output power of the hub motor as a generator is defined as: ,in This is the output voltage after motor rectification. This is the charging current; MPPT algorithm iteration process: Measure current power ; Apply a small perturbation to the pulse width modulation duty cycle (For example, +0.5%) Measure the power after disturbance ; Calculate power change ; like If so, continue the perturbation in the same direction. Then the disturbance in the opposite direction ( ); Repeat the above steps every 50 milliseconds until... ( (This is a preset convergence threshold), which means that it is considered to have converged to the maximum power point at the current speed. Once the braking signal is confirmed, the MPPT function is shielded by the hardware level signal, and the bidirectional DC-DC converter immediately switches to the regenerative torque command mode. Braking torque coordination and distribution in braking intervention state: Upon receiving a dual-path redundant braking signal that has been confirmed by de-jittering, the state machine forcibly jumps to the braking intervention state and executes the following sub-steps: Braking signal authenticity verification: Read the four signals from the two independent Hall sensors inside the left and right brake levers. Only when both sensors on each lever output valid levels and the difference between the two signals is less than a preset error threshold (e.g., 5%), is it confirmed as a genuine braking signal. Otherwise, it is determined to be a sensor malfunction, only the malfunction indicator light is illuminated, and no braking intervention is performed. Coordination of regenerative braking torque and mechanical braking: In the braking intervention state, the MCU performs two operations simultaneously: cutting off the assist output of the hub motor (PWM output set to zero), and adjusting the real-time vehicle speed v and brake travel signal. (0~1, 0 means no braking, 1 means full braking) Calculate the target value of regenerative braking torque; First, calculate the basic regeneration torque: ,in This is the maximum regenerative braking torque coefficient. The effective speed limit for regenerative braking (e.g., 15 m / s); Then slip ratio control is introduced, based on the vehicle speed v and wheel speed. (Calculate slip ratio based on motor Hall signal) ,in To prevent division by zero, the final regenerative braking torque is: ,in, The optimal slip ratio is (e.g., 0.15). This represents the hysteresis band (e.g., 0.05). When the slip ratio enters ( , When the interval is ) The speed decreases linearly to zero. Furthermore, when the vehicle speed is below a preset low-speed exit threshold (e.g., 1.5 m / s), The brakes are forced to zero, relying entirely on mechanical braking to stop the vehicle. Downhill braking compensation: If the slope angle sensor detects a downhill slope, the upper limit of the regenerative braking torque is linearly reduced. For example, when the slope angle is -10°, the upper limit of the regenerative braking torque is reduced to 0; State Exit Transition and Recovery: After the braking signal disappears, the state machine first transitions from the braking intervention state to the braking exit transition state. In this transition state, the regenerative braking torque... It decays linearly to zero within 0.5 seconds: After the decay is complete, the generated cycling difficulty index R and hysteresis comparison results are reread, and the next state is determined according to the following rules: If R> Then it enters the assisted establishment state; If R < Then it enters the recycling establishment state; Otherwise, it enters a state of complete human idleness; When it is necessary to exit from the assist or regenerative braking state due to changes in conditions, it first enters the corresponding exit transition state (assist exit transition state or regenerative braking exit transition state), linearly decaying the auxiliary torque or regenerative power to zero within 0.5 seconds, and finally disengaging the one-way overrunning clutch or stopping the MPPT, ensuring that there is no sudden torque change during the entire switching process.

[0019] In one embodiment, the MCU computing controller 4 generates a cycling difficulty index through the weighted fusion filter. The weighted fusion filter is a Kalman filter, which dynamically weights and fuses the real-time slope angle signal output by the slope angle sensor 6, the pedal frequency signal output by the cadence sensor 5, and the real-time vehicle speed signal output by the vehicle speed Hall sensor 7 to output a continuous cycling difficulty index. The Schmidt hysteresis comparator sets state transition thresholds for the riding difficulty index. The state transition thresholds include a first threshold for entering the assist mode, a second threshold for exiting the assist mode, a third threshold for entering the recovery mode, and a fourth threshold for exiting the recovery mode. The first threshold for entering the assist mode is higher than the second threshold for exiting the assist mode, and the third threshold for entering the recovery mode is lower than the fourth threshold for exiting the recovery mode. All state transitions must satisfy a corresponding duration greater than the preset anti-shake delay.

[0020] In one embodiment, the braking torque coordination algorithm module performs the following operations during the braking intervention state: The system receives two Hall signals from the dual-redundant brake lever brake sensor 8 and only confirms a real braking signal when both signals are valid at the same time and the difference between the two signals is less than a preset error threshold. The wheel slip ratio is calculated based on the real-time vehicle speed measured by the vehicle speed Hall sensor 7. When the wheel slip ratio is less than the preset optimal slip ratio threshold, the regenerative braking torque is allowed to increase linearly with the braking stroke. When the slip ratio is greater than the preset optimal slip ratio threshold, the regenerative braking torque is immediately set to zero and mechanical braking is completely relied upon. The upper limit of the regenerative braking torque decreases linearly as the vehicle speed decreases. When the vehicle speed is lower than the preset low-speed exit threshold, the regenerative braking is completely disengaged. Furthermore, under downhill braking conditions, the maximum limit of the regenerative braking torque is linearly reduced based on the negative slope value of the slope angle sensor.

[0021] In one embodiment, the pre-stored two-dimensional torque curve mapping table uses the slope angle signal and cadence signal as two-dimensional input variables and the basic auxiliary torque as the output variable. The pre-stored two-dimensional torque curve mapping table is stored in the non-volatile memory inside the MCU operation controller 4 and contains torque values ​​of 25 grid points, including no less than 5 slope angle discrete points and 5 cadence discrete points. The real-time torque is calculated by bilinear interpolation between adjacent grid points. The multiplicative correction coefficient for rider fatigue factor is obtained by normalizing the fatigue signal collected by the cadence sensor 5. Its default value is 1.0, and its range is limited to between 0.8 and 1.5. Furthermore, the rate of change of this correction coefficient is limited by a first-order low-pass filter.

[0022] In one embodiment, the bidirectional DC-DC converter activates perturbation-observation maximum power point tracking only when the multimodal finite state machine is in the recovery establishment state or steady-state recovery state and has not received any braking signal; After activation, the bidirectional DC-DC converter applies a small perturbation to the pulse width modulation duty cycle every 50 milliseconds and observes the change in the input power output of the booster hub motor 2 as a generator. If the power increases, it continues to perturb in the same direction; if the power decreases, it perturbs in the opposite direction. After multiple iterations of three to five times, it converges to the maximum power point at the current speed. Once the braking signal is confirmed, the maximum power point tracking using the disturbance observation method is immediately shielded by the hardware-level signal, and the bidirectional DC-DC converter switches to the regenerative torque command mode directly given by the braking torque coordination algorithm module.

[0023] In one embodiment, the multimodal finite state machine is further provided with an exit transition state, which includes an assisted exit transition state, a recovery exit transition state, and a braking exit transition state. When the conditions for exiting from the steady-state assisted state to the pure manual idle state are met, the multimodal finite state machine first jumps to the assisted exit transition state. In this state, the auxiliary torque output by the assisted hub motor 2 linearly decays to 0 within 0.5 seconds before the one-way overrunning clutch is disengaged. When the braking signal disappears, the multimodal finite state machine first jumps from the braking intervention state to the braking exit transition state, linearly decays the regenerative braking torque to 0 within 0.5 seconds, and then decides to enter the recovery establishment state, the assist establishment state, or directly enter the pure human idle state based on the current riding difficulty index and the hysteresis comparator result.

[0024] In one embodiment, the dual-redundant brake lever brake sensors 8 are respectively installed on the left and right brake levers 10 of the bicycle. Each brake lever 10 is provided with two independent Hall effect sensors from different sources. The two signals are respectively input to the MCU computing controller 4 through different analog-to-digital conversion channels.

[0025] In one embodiment, the MCU operational controller 4 is equipped with a hardware de-jitter filter and a software de-jitter algorithm. It continuously samples 10 times, and only if 8 or more of them are valid levels is it confirmed as a real braking signal. If only one sensor outputs a valid level and the duration is longer than the preset fault confirmation time, it is determined to be a sensor fault. At this time, the MCU operational controller 4 illuminates the fault indicator light on the handlebar control panel 11 without entering the braking intervention state. The handlebar control panel 11 is located above the handlebars.

[0026] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A lightweight, non-restrictive, bidirectional kinetic energy recovery self-powered bicycle system, comprising a bicycle (1), characterized in that, The bicycle (1) includes: Power-assisted hub motor (2): It integrates a one-way overrunning clutch (3), which is located between the motor rotor and the wheel drive unit and is used to mechanically cut off the power transmission path when the power-assisted hub motor (2) is not powered. MCU operation controller (4): It has a multimodal finite state machine, a weighted fusion filter and a braking torque coordination algorithm module embedded in it, and is electrically connected to the power wheel hub motor (2). At the same time, it receives real-time acquisition signals from the pedal frequency sensor (5), the slope angle sensor (6), the vehicle speed Hall sensor (7) and the dual-redundant brake handle braking sensor (8). Energy storage battery pack (9): It has an integrated battery management system, which includes a battery temperature monitoring module and a charge / discharge overcurrent protection module. The energy storage battery pack (9) forms a charge / discharge closed loop with the power wheel hub motor (2) through a bidirectional DC converter. The MCU operation controller (4) generates a cycling difficulty index from the collected signals, and drives the multimodal finite state machine to switch between pure human idle state, assist establishment state, steady-state assist state, recovery establishment state, steady-state recovery state and braking intervention state based on the cycling difficulty index and combined with the Schmidt hysteresis comparator. When in the boost-up state or steady-state boost state, the MCU operation controller (4) outputs auxiliary torque to the boost hub motor (2). When the bidirectional DC-DC converter is in the recovery establishment state or steady-state recovery state and no braking signal is received, it performs maximum power point tracking using the perturbation-observation method. When the braking signal is received from the dual-redundant brake lever brake sensor (8), the multimodal finite state machine is forced to jump to the braking intervention state with the highest priority. The braking torque coordination algorithm module regenerates the braking torque by linear mapping according to the real-time vehicle speed signal and the brake stroke signal, and at the same time cuts off the power assist output of the power assist hub motor (2) and shields the maximum power point tracking.

2. The lightweight, non-restrictive, bidirectional kinetic energy recovery self-powered bicycle system according to claim 1, characterized in that: The MCU operation controller (4) generates a cycling difficulty index through the weighted fusion filter. The weighted fusion filter is a Kalman filter. The Kalman filter dynamically weights and fuses the real-time slope angle signal output by the slope angle sensor (6), the pedal frequency signal output by the pedal frequency sensor (5), and the real-time vehicle speed signal output by the vehicle speed Hall sensor (7) to output the cycling difficulty index. The Schmidt hysteresis comparator sets state transition thresholds for the cycling difficulty index. The state transition thresholds include a first threshold for entering the assist mode, a second threshold for exiting the assist mode, a third threshold for entering the recovery mode, and a fourth threshold for exiting the recovery mode. The first threshold for entering the assist mode is higher than the second threshold for exiting the assist mode, and the third threshold for entering the recovery mode is lower than the fourth threshold for exiting the recovery mode.

3. The lightweight, non-restrictive, bidirectional kinetic energy recovery self-powered bicycle system according to claim 1, characterized in that: The braking torque coordination algorithm module performs the following operations during the braking intervention state: The system receives two Hall signals from the dual-redundant brake lever brake sensor (8), and only when both signals are valid at the same time and the difference between the two signals is less than a preset error threshold is it confirmed as a real braking signal. The wheel slip ratio is calculated based on the real-time vehicle speed measured by the vehicle speed Hall sensor (7). When the wheel slip ratio is less than the preset optimal slip ratio threshold, the regenerative braking torque is allowed to increase linearly with the braking stroke. When the slip ratio is greater than the preset optimal slip ratio threshold, the regenerative braking torque is immediately set to zero. The upper limit of the regenerative braking torque decreases linearly as the vehicle speed decreases, and regenerative braking is completely disengaged when the vehicle speed is lower than the preset low-speed exit threshold.

4. The lightweight, non-restrictive, bidirectional kinetic energy recovery self-powered bicycle system according to claim 1, characterized in that: The MCU operation controller (4) outputs the auxiliary torque smoothed by a first-order low-pass filter to the power-assisted hub motor (2) according to the pre-stored two-dimensional torque curve mapping table and the rider fatigue factor multiplicative correction coefficient. The pre-stored two-dimensional torque curve mapping table uses the slope angle signal and the pedal frequency signal as two-dimensional input variables and the basic auxiliary torque as the output variable. The pre-stored two-dimensional torque curve mapping table is stored in the non-volatile memory inside the MCU operation controller (4) and contains the torque values ​​of 25 grid points, including no less than 5 slope angle discrete points and 5 pedal frequency discrete points. The real-time torque is calculated by bilinear interpolation between adjacent grid points. The multiplicative correction coefficient for rider fatigue factor is obtained by normalizing the fatigue signal collected by the cadence sensor (5), with a default value of 1.0 and a range limited to 0.8 to 1.

5.

5. The lightweight, non-restrictive, bidirectional kinetic energy recovery self-powered bicycle system according to claim 1, characterized in that: The bidirectional DC-DC converter activates perturbation-observation maximum power point tracking only when the multimodal finite state machine is in the recovery establishment state or steady-state recovery state and has not received any braking signal; After activation, the bidirectional DC-DC converter applies a small perturbation to the pulse width modulation duty cycle every 50 milliseconds and observes the change in the input power output of the power-assisted hub motor (2) as a generator. If the power increases, it continues to perturb in the same direction; if the power decreases, it perturbs in the opposite direction. After multiple iterations, it converges to the maximum power point at the current speed. Once the braking signal is confirmed, the maximum power point tracking using the disturbance observation method is immediately shielded by the signal, and the bidirectional DC-DC converter switches to the regenerative torque command mode directly given by the braking torque coordination algorithm module.

6. The lightweight, non-restrictive, bidirectional kinetic energy recovery self-powered bicycle system according to claim 1, characterized in that: The multimodal finite state machine also includes exit transition states, which include assist exit transition states, recovery exit transition states, and braking exit transition states. When the conditions for exiting from the steady-state assisted state to the pure human idle state are met, the multimodal finite state machine first jumps to the assisted exit transition state. In this state, the auxiliary torque output by the assisted hub motor (2) linearly decays to 0 within 0.5 seconds before the one-way overrunning clutch is disengaged. When the braking signal disappears, the multimodal finite state machine first jumps from the braking intervention state to the braking exit transition state, linearly decays the regenerative braking torque to 0 within 0.5 seconds, and then decides to enter the recovery establishment state, the assist establishment state, or directly enter the pure human idle state based on the current riding difficulty index and the hysteresis comparator result.

7. The lightweight, non-restrictive, bidirectional kinetic energy recovery self-powered bicycle system according to claim 1, characterized in that: The dual-redundant brake lever brake sensors (8) are respectively installed on the left and right brake levers (10) of the bicycle. Each brake lever (10) is equipped with two independent Hall effect sensors from different sources. The two signals are respectively input to the MCU operation controller (4) through different analog-to-digital conversion channels.

8. The lightweight, non-restrictive, bidirectional kinetic energy recovery self-powered bicycle system according to claim 1, characterized in that: The MCU operation controller (4) samples 10 times continuously. Only when 8 or more of the samples are valid can the signal be confirmed as a real braking signal. If only one sensor outputs a valid level and the duration is longer than the preset fault confirmation time, the sensor is determined to be faulty. At this time, the MCU operation controller (4) illuminates the fault indicator on the handlebar control panel (11) without entering the braking intervention state. The handlebar control panel (11) is located above the handlebar.

Images