A control method for a two-wheeled electric bicycle
The dual-wheeled electric vehicle control method, which uses multi-source information for comprehensive decision-making, solves the problems of insufficient terrain perception, poor turning safety, and misjudgment of anti-skid in existing technologies. It achieves smooth assistance and stable traction on different road surfaces, improving handling safety and terrain adaptability.
Patent Information
- Application Number
- CN202611323808.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-28
- Publication Date
- 2026-11-17
AI Technical Summary
Existing two-wheeled electric vehicle control solutions lack terrain perception and adaptive capabilities, cannot distinguish the control requirements of different road surfaces, lack effective safety protection mechanisms when turning, and have difficulty distinguishing between turning differential and slip differential in anti-slip control. Furthermore, the various control function modules lack system-level collaborative integration.
By acquiring information on human-machine interaction forces, the motion status of the left and right drive wheels, the vehicle's pitch and yaw attitude, and terrain patterns, and combining multi-source information under the same control framework, comprehensive decision-making is carried out, including torque fusion on hard surfaces, speed closed-loop control on soft surfaces, and anti-slip control. Anti-slip judgment is made by using yaw rate correction and wheel speed change rate.
It achieves smooth power assistance on hard surfaces and stable traction on soft surfaces, and ensures safety through yaw correction when turning. It solves the problems of poor terrain adaptability and easy misjudgment of anti-skid, without adding an extra vehicle speed sensor.
Smart Images

Figure CN123104838A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of two-wheeled electric vehicle control technology, and more particularly to a two-wheeled electric vehicle control method. Background Technology
[0002] Two-wheeled electric vehicles are low-speed electric vehicles that are assisted by the operator pushing or pulling to trigger an electric motor. Common forms include electric handcarts, electric cargo camping vehicles, electric ATVs, and warehousing and logistics transport vehicles. These vehicles typically have independent drive wheels on the left and right sides. They determine the operator's intention to move by detecting the pushing or pulling force applied to the handlebars and control the left and right drive motors to output corresponding assist torque. Compared to single-wheel drive systems, they have stronger load-bearing capacity and terrain adaptability.
[0003] Currently, the most mainstream control scheme for two-wheeled electric bicycles is a proportional control scheme based on human-machine interaction. This scheme places force sensors (such as strain gauges or tension / compression sensors) at the handlebars to collect the operator's pushing and pulling force signals in real time, and maps them to motor torque according to a preset fixed ratio or fixed gain, outputting them to the left and right drive motors respectively. Building upon this, some improved schemes further introduce closed-loop speed control to maintain the target speed, or introduce inertial measurement units to detect pitch / yaw rates for slope compensation or turning speed limiting.
[0004] However, the above solution still has the following shortcomings:
[0005] First, it lacks terrain awareness and adaptive capabilities, failing to differentiate control requirements across different road surfaces. The core logic of proportional control schemes is to directly map the operator's applied force into motor torque, meaning the motor outputs the same amount of torque as the operator pushes. This logic is generally feasible on hard, flat surfaces, but on soft surfaces (such as sand), the same torque output may not be enough to start the vehicle (insufficient torque leading to getting stuck), or it may cause the wheels to spin continuously and sink (excessive torque leading to digging). While speed-closed-loop schemes can maintain the target speed, the operator will clearly feel the vehicle moving forward uncontrollably, reducing the direct sense of control.
[0006] Secondly, there is a lack of effective safety protection mechanisms during cornering. Proportional control schemes continue to increase assist as force is applied during cornering, which can easily cause the vehicle to lurch forward before entering the corner or result in excessive drive force on the outer wheel, leading to rollover. While some schemes introduce yaw rate for speed limiting, this is merely a simple speed limit and lacks an active safety mechanism to lock the torque limit before entering the corner, thus failing to fundamentally prevent the assist from continuing to increase after entering the corner.
[0007] Third, anti-slip control on soft surfaces presents problems of misjudgment and cost. Existing solutions mostly determine slippage by detecting the speed difference between the left and right wheels; that is, when one wheel's speed is significantly higher than the other, it is considered slipping, and its torque is reduced. However, when a vehicle is actually turning, the outer wheel will inevitably turn faster than the inner wheel—this is a normal physical speed difference. Existing solutions cannot distinguish between "turning differential" and "slip differential," therefore, it is very easy to misjudge normal differential as slippage and incorrectly reduce torque, causing a loss of turning power and a very poor user experience. Another solution, which relies on an additional vehicle speed sensor to calculate the slip ratio, can improve the accuracy of the judgment, but it increases the system cost and complexity.
[0008] Fourth, the various control modules operate independently, lacking system-level coordination and integration. Modules such as power assist control, speed closed-loop control, attitude compensation, and anti-slip control are typically implemented using independent algorithms, resulting in a lack of information sharing and collaborative decision-making. For example, in a turning scenario, the anti-slip module only determines slippage based on wheel speed differences, failing to recognize when the vehicle is turning. Consequently, it may misinterpret the normal high speed of the outer wheel as slippage and incorrectly reduce torque. Similarly, in a bumpy scenario, the attitude compensation module only outputs ramp compensation torque based on pitch angle, unable to determine if the road surface is bumpy. Even when the pitch angle signal is contaminated by vibration, it continues to output compensation torque, causing unnecessary acceleration and deceleration. The root cause of these problems lies in the independent operation of each module, the lack of unified scheduling, and the inability to comprehensively utilize multi-source information for integrated decision-making within a unified framework.
[0009] Therefore, a unified solution is needed that can switch control targets according to road surface characteristics and integrate human-machine interaction forces, wheel speed and inertia information to complete power assist output, turning torque limiting, traction on soft roads and anti-skid control under the same control framework. Summary of the Invention
[0010] To address the problems existing in the background technology, this invention proposes a control method for a two-wheeled electric vehicle, aiming to solve the problems of existing two-wheeled electric vehicles lacking adaptive control capabilities based on road surface differences, resulting in poor smoothness on hard surfaces, insufficient traction and easy slippage on soft surfaces, lack of effective safety protection mechanisms under dynamic conditions such as turning, and difficulty in distinguishing between turning differential and slippage differential in anti-slip control.
[0011] To solve the above-mentioned technical problems, the present invention provides a control method for a two-wheeled electric bicycle, comprising the following steps:
[0012] The system acquires information on the human-machine interaction force and grip state applied by the operator, the motion state information of the left and right drive wheels, the vehicle pitch attitude information and the vehicle yaw attitude information, and the terrain pattern information used to characterize the road surface adhesion conditions. The terrain pattern information includes at least a first mode and a second mode corresponding to hard road surface and soft road surface, respectively.
[0013] Based on the gripping status information, it is determined whether the operator is in a gripping state. If the operator is not in a gripping state, the output of drive torque to the left and right drive wheels is terminated.
[0014] When the vehicle is in a gripping state and the terrain mode information is the first mode, a first assist component that changes with vehicle speed is determined based on the motion state information, and a second assist component that changes with applied force is determined based on the human-machine interaction force information. A basic drive requirement is generated based on the first assist component and the second assist component. Then, based on the pitch attitude information and the yaw attitude information, the basic drive requirement is adjusted for slope adaptability and cornering safety, respectively, to generate a first torque command.
[0015] When the machine is in a gripping state and the terrain mode information is the second mode, the target motion speed and feedforward drive component are determined according to the human-machine interaction force information. The target motion speed is independently controlled by closed-loop following according to the motion state information, and closed-loop adjustment components of each wheel are generated. The initial torque of each wheel is obtained according to the feedforward drive component and the closed-loop adjustment component. Then, a first anti-slip intervention is performed based on the rate of change of the motion state of each drive wheel, and a second anti-slip intervention is performed according to the difference in motion state between the left and right drive wheels. The activation condition of the second anti-slip intervention is constrained by the yaw attitude information. The initial torque is corrected according to the results of the first and second anti-slip interventions, and a second torque command is generated.
[0016] The first torque command or the second torque command is output to the left drive wheel and the right drive wheel, respectively.
[0017] In some embodiments, the step of terminating the output of drive torque to the left and right drive wheels includes:
[0018] Execute braking action until the absolute value of the driving speed determined based on the motion state information is lower than the preset stopping speed;
[0019] After coming to a stop, the absolute value of the pitch attitude information is used to determine: if the absolute value exceeds a first threshold, the braking lock state is maintained; if the absolute value is lower than a second threshold, the braking is released and the vehicle is allowed to move freely; if the absolute value is between the second threshold and the first threshold, the current state is maintained.
[0020] In some embodiments, before determining the first assist component that varies with vehicle speed based on the motion state information, the method further includes:
[0021] Determine the current vehicle speed based on the motion state information;
[0022] Based on the current vehicle speed, force dead zone thresholds corresponding to the operator's pushing and pulling forces are determined respectively. The force dead zone threshold when the current vehicle speed is zero or lower than the preset low speed threshold is greater than the force dead zone threshold when the current vehicle speed is higher than the preset travel speed threshold.
[0023] When the amplitude of the human-machine interaction force information does not exceed the corresponding dead zone threshold of the interaction force, it is determined that there is no effective force application intention.
[0024] In some implementations, the step of generating basic drive requirements based on the first assist component and the second assist component includes:
[0025] Determine the current vehicle speed based on the motion state information;
[0026] The current vehicle speed is converted into the first assist component according to a preset segmented mapping curve. The segmented mapping curve includes at least: a zero assist section when the vehicle speed is below a first threshold, an increasing section when the vehicle speed is between the first threshold and the second threshold, a maintenance section when the vehicle speed is between the second threshold and the third threshold, and a decreasing or limiting section after the vehicle speed exceeds the third threshold.
[0027] The human-machine interaction force information is normalized after deducting the corresponding dead zone threshold, and the normalized value is converted into the second assist component through a preset nonlinear mapping.
[0028] The first assist component and the second assist component are added together to obtain the basic drive requirement.
[0029] In some embodiments, the slope adaptation adjustment step includes:
[0030] Obtain the bump status information of the vehicle body;
[0031] When the vehicle is determined to be in a rough driving state based on the bumpy state information, the gain of the slope adaptive adjustment is limited or reduced.
[0032] When it is determined from the bumpy state information that the vehicle body is not in a rough driving state and the pitch attitude information exceeds the preset slope dead zone, the slope gain is calculated based on the magnitude of the pitch attitude information exceeding the slope dead zone, and the basic drive requirements are adjusted using the slope gain.
[0033] In some implementations, the turning safety adjustment step includes:
[0034] When the absolute value of the yaw attitude information exceeds the preset cornering threshold, the current value or the previous cycle value of the basic drive requirement before entering the turning state is locked as the torque upper limit.
[0035] During the turning state, the basic drive demand after the slope adaptive adjustment is limited to no more than the torque limit, and the drive demand is further attenuated according to the absolute value of the yaw attitude information, wherein the larger the absolute value of the yaw attitude information, the greater the attenuation.
[0036] When the absolute value of the yaw attitude information is lower than the preset corner exit threshold, the turning state is released and the torque limit lock is canceled.
[0037] In some embodiments, before determining the target motion velocity and feedforward drive component based on the human-machine interaction force information, the method further includes:
[0038] The direction of the human-machine interaction force information is determined;
[0039] If the direction of the human-machine interaction force in the current cycle is opposite to the direction of the historical filtered force, the system enters the reversal confirmation window. During the reversal confirmation window, the historical filtered force is attenuated with the goal of zero while maintaining its original sign.
[0040] The target motion speed in the opposite direction is only allowed to be established after the human-machine interaction force in the opposite direction has been continuously maintained for a preset number of confirmation frames.
[0041] In some embodiments, after obtaining the initial torque of each wheel based on the feedforward drive component and the closed-loop adjustment component, the method further includes:
[0042] The maximum allowable torque for each wheel is initially set to the first maximum torque.
[0043] When the speed of any drive wheel is lower than the preset speed threshold for getting out of trouble, the allowable torque limit of that drive wheel is temporarily switched to a second torque limit that is greater than the first torque limit.
[0044] When the rotational speed of the drive wheel exceeds the escape speed threshold, the allowable torque limit of the drive wheel is restored to the first torque limit.
[0045] The modification of the initial torque specifically includes: limiting the initial torque to no higher than the current allowable torque limit of the corresponding drive wheel.
[0046] In some embodiments, the step of performing the first anti-slip intervention based on the rate of change of the motion state of each drive wheel includes:
[0047] The angular acceleration of each drive wheel is determined based on the difference between the current cycle speed and the previous cycle speed.
[0048] The angular acceleration is filtered.
[0049] When the filtered angular acceleration of any drive wheel continuously exceeds a preset acceleration threshold in the direction of the target motion velocity, the drive wheel is determined to be slipping.
[0050] When the slippage state is entered, the closed-loop control integral term of the drive wheel is reduced, and the output torque of the drive wheel is reduced by a preset penalty ratio during the slippage period.
[0051] In some embodiments, the step of performing the second anti-slip intervention based on the difference in motion states between the left and right drive wheels includes:
[0052] When the absolute value of the yaw attitude information is lower than the preset turning threshold, the speed difference between the left drive wheel and the right drive wheel is calculated.
[0053] When the absolute value of the speed difference exceeds a preset differential threshold, the drive wheel with the higher speed is identified as the slipping wheel; a penalty coefficient is determined based on the magnitude by which the speed difference exceeds the differential threshold, and the penalty coefficient is used to reduce the output torque of the slipping wheel and the closed-loop control integral term;
[0054] When the absolute value of the yaw attitude information reaches or exceeds the turning threshold, the judgment of the speed difference in this cycle is skipped.
[0055] The beneficial effects of this invention are as follows: By acquiring the terrain pattern and switching the control target accordingly, on hard surfaces, the first assist component mapped by vehicle speed and the second assist component mapped by force are fused together, and the torque command is output after yaw rate correction. On soft surfaces, the force is mapped to the target speed and wheel speed is closed-loop. At the same time, dual-criteria anti-skid is performed based on the wheel speed change rate and the difference between the left and right wheel speeds. The second criterion is constrained by the yaw rate to avoid misjudgment of turning. Thus, under the same control framework, human-machine interaction force, wheel speed and inertia information are integrated, achieving smooth assist on hard surfaces and stable traction on soft surfaces. Safety is ensured by yaw correction when turning. At the same time, turning differential and slip differential are distinguished without adding an additional vehicle speed sensor, solving the problems of poor terrain adaptability and easy misjudgment of anti-skid in the existing solution. Attached Figure Description
[0056] Figure 1 This is a flowchart illustrating the two-wheeled electric vehicle control method disclosed in an embodiment of the present invention. Detailed Implementation
[0057] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the content of this invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to this invention are shown in the accompanying drawings, not all of them.
[0058] This embodiment proposes a control method for a two-wheeled electric bicycle, such as... Figure 1 As shown, it includes the following steps:
[0059] Step 1, Data Acquisition: Acquire information on the human-machine interaction force and grip state applied by the operator, motion state information of the left and right drive wheels, vehicle pitch attitude information and vehicle yaw attitude information, and acquire terrain pattern information used to characterize road surface adhesion conditions. The terrain pattern information includes at least a first mode and a second mode corresponding to hard road surface and soft road surface, respectively.
[0060] Step 1 of this solution acquires information from multiple sensors, providing a complete data foundation for subsequent power assist control. Specifically, human-machine interaction force information is used to identify the operator's pushing and pulling intentions; grip status information is used to determine whether the operator is controlling the vehicle; left and right wheel speed information is used to calculate vehicle speed and wheel speed difference; pitch attitude information is used to identify slopes; yaw attitude information is used to identify turns; and terrain mode information is used to switch between different control strategies. Through the fusion of the above multi-dimensional information, the vehicle can adaptively adjust its power assist characteristics according to different road conditions and operator intentions, laying the data foundation for subsequent torque fusion control on hard surfaces and speed closed-loop control on soft surfaces.
[0061] In a preferred embodiment, human-machine interaction force information is acquired through a human-machine interaction force detection unit located at the vehicle's grip. This detection unit can be a strain gauge, pressure sensor, tension / compression sensor, Hall displacement sensor, or other sensor capable of characterizing push / pull intentions. Vehicle pitch and yaw attitude information is acquired through an inertial measurement unit (IMU), which can be a six-axis or nine-axis IMU. Terrain mode information can be manually selected by the user via physical buttons, a touchscreen, or a mobile terminal, or it can be automatically identified by the controller based on wheel speed fluctuations, wheel angular acceleration, left / right wheel speed differences, attitude vibrations, or drive loads; no limitation is made here.
[0062] Step 2, Grip Status Judgment: Determine whether the operator is in a gripping state based on the grip status information. If the operator is not in a gripping state, terminate the output of drive torque to the left and right drive wheels.
[0063] Step 2 primarily achieves safe control of vehicle drive enablement through the gripping status. When the operator is not holding the vehicle, the drive torque output is automatically cut off to prevent the vehicle from moving on its own and causing safety hazards. Its core principle is to make "human gripping" a necessary condition for drive enablement, ensuring safety while also providing a trigger condition for subsequent release and parking control.
[0064] Specifically, when not in a holding state, the above termination steps include:
[0065] Step 201: Perform braking action until the absolute value of the driving speed determined based on the motion state information is lower than the preset stopping speed;
[0066] Step 202: After stopping, determine the absolute value of the pitch attitude information: if the absolute value exceeds the first threshold, maintain the braking lock state; if the absolute value is lower than the second threshold, release the brake and allow the vehicle to move freely; if the absolute value is between the second threshold and the first threshold, maintain the current state.
[0067] The aforementioned termination steps achieve parking control through a two-stage safety assessment after the brakes are released. The first stage is the braking stage, which actively brakes to rapidly decelerate the vehicle to a stop, preventing it from continuing to roll due to inertia after the brakes are released. The second stage is the angle judgment stage, which determines the parking strategy based on the slope. For example, on a slope, the brakes are locked to prevent the vehicle from rolling, while on flat ground, the brakes are released to allow the vehicle to move freely. The hysteresis range formed between the first and second thresholds avoids the vehicle frequently switching between locking and releasing states near critical slopes, improving control stability.
[0068] In one example, the grip signal in step 2 employs asymmetric anti-shake processing: when a grip is detected, it is immediately confirmed as valid, enabling rapid connection; when the grip signal disappears, it must disappear continuously for a preset period (e.g., 5 control cycles) before release is confirmed. This processing method ensures both timely grip response and reduces false alarms due to transient power outages caused by bumpy roads.
[0069] In one example, during the angle determination phase of step 202 above, if the wheel speed exceeds a preset rollback safety threshold, the system immediately reverts to the braking phase and re-executes the braking action. This reversal mechanism effectively prevents accidental rollback caused by brake failure or external disturbance after parking on a slope.
[0070] Step 3, Hard Surface Adjustment: When in the grip state and the terrain mode information is in the first mode, the first assist component that changes with vehicle speed is determined based on the motion state information, and the second assist component that changes with applied force is determined based on the human-machine interaction force information. The basic drive demand is generated based on the first and second assist components. Then, the basic drive demand is adjusted for slope adaptability and cornering safety based on pitch attitude information and yaw attitude information, respectively, and the first torque command is generated.
[0071] Step 3 of this solution primarily controls the two wheels on hard surfaces through dual-track torque fusion. Specifically, the first assist component varies with vehicle speed, providing basic adaptive assistance to maintain stable power output during driving. The second assist component varies with the operator's applied force, reflecting the operator's real-time acceleration or deceleration intentions. After fusion, pitch attitude information is used for slope compensation, and yaw attitude information is used for cornering torque limiting, ultimately forming a smooth and safe driving torque. This solution ensures assist tracking during normal driving while providing active safety protection in special conditions such as slopes and cornering, solving problems such as insufficient power on slopes, excessive assistance on downhill slopes, and forward thrust during cornering with traditional fixed-ratio assist.
[0072] In a preferred embodiment, before determining the first assist component, the following is also included:
[0073] Step 301: Determine the current vehicle speed based on the motion status information;
[0074] Step 302: Determine the dead zone thresholds of the force corresponding to the operator's pushing and pulling forces based on the current vehicle speed. The dead zone threshold for the force when the current vehicle speed is zero or lower than the preset low speed threshold is greater than the dead zone threshold for the force when the current vehicle speed is higher than the preset travel speed threshold.
[0075] Step 303: When the amplitude of the human-machine interaction force information does not exceed the corresponding dead zone threshold, it is determined that there is no valid force application intention. It should be noted that determining no valid force application intention in step 303 only means that the human-machine interaction force information is not used as valid input in the calculation. At this time, the second assist component is zero, but the vehicle can still obtain basic follow-up assist based on the vehicle speed through the first assist component. In other words, the determination in step 303 will not interrupt the calculation of basic drive requirements in subsequent steps 304-307; it only affects the value of the second assist component.
[0076] Steps 301-303 above aim to achieve a balance between preventing accidental touches when stationary and maintaining high sensitivity while in motion through a dynamic dead zone correlated with vehicle speed. Specifically, a larger dead zone threshold is used when the vehicle is stationary or at extremely low speeds, so that slight, unintentional touches by the operator will not trigger the power steering, preventing sudden lurches in the vehicle. When the vehicle is moving, it automatically switches to a smaller dead zone threshold, so that a slight force from the operator will elicit a response, ensuring responsiveness. At the same time, independent dead zone thresholds are used for the push and pull directions, which can adapt to the differences in the physical characteristics of the vehicle structure in the push and pull directions, further improving the finesse of the control.
[0077] Based on steps 301-303 above, in one example, when the vehicle speed gradually increases from zero to a preset travel speed threshold (e.g., 0.5 m / s), the dead zone threshold decreases linearly from a stationary value (e.g., 15 N) to a traveling value (e.g., 8 N), achieving a smooth transition of the dead zone and avoiding torque jumps caused by sudden changes in the dead zone.
[0078] In a preferred embodiment, the calculation steps for the basic driving requirements in this embodiment include:
[0079] Step 304: Determine the current vehicle speed based on the motion status information;
[0080] Step 305: Convert the current vehicle speed into a first assist component according to a preset segmented mapping curve. The segmented mapping curve includes at least: a zero assist section when the vehicle speed is below a first threshold, an increasing section when the vehicle speed is between the first threshold and the second threshold, a maintenance section when the vehicle speed is between the second threshold and the third threshold, and a decreasing or limiting section after the vehicle speed exceeds the third threshold.
[0081] Step 306: After deducting the corresponding dead zone threshold from the human-machine interaction force information, normalize the information and convert the normalized value into a second assist component through a preset nonlinear mapping.
[0082] Step 307: Add the first assist component and the second assist component to obtain the basic drive requirements.
[0083] Steps 304-307 above demonstrate the specific implementation of dual-track torque fusion. Specifically, the segmented mapping from vehicle speed to the first assist component includes four segments: zero assist, increasing, maintaining, and decreasing or limiting. The zero assist segment eliminates the creeping sensation at extremely low speeds; the increasing segment provides smooth acceleration from a standstill; the maintaining segment ensures stable assist within the commonly used speed range; and the decreasing or limiting segment automatically narrows the output to prevent speeding when approaching the speed limit, forming a smooth speed management mode. The human-machine interaction force, after dead zone subtraction, normalization, and nonlinear mapping, is converted into the second assist component. Its nonlinear mapping makes changes sensitive in the small force range and smooth in the large force range, providing a delicate feel and preventing the mechanics from reaching physical limits. The first and second assist components are added and merged to form a unified system of continuous, speed-dependent assist and responsive assist that reflects the operator's immediate force requirements.
[0084] In one example, the threshold values for each segment of the segmented mapping curve in step 305 can be calibrated as follows: the speed dead zone is 0.1 m / s (below this speed, the base torque is zero), the peak starting speed is 0.3 m / s (the base torque increases linearly from zero), the peak ending speed is 0.8 m / s (the base torque reaches its peak and remains there), and the speed limit speed is 2.0 m / s (beyond this speed, the base torque gradually decreases to its maintenance value). These values are merely examples; actual calibration values can be adjusted according to the vehicle model and motor characteristics.
[0085] In one example, the nonlinear mapping in step 305 uses a cubic smooth step function s(y) = y² × (3 - 2y), where y is the normalized effective force (range 0 to 1). This function is approximately linear in the low force range to ensure the predictability of small operations, and automatically smooths saturation in the high force range, so that the second assist component gradually approaches the upper limit as the applied force increases, avoiding the abrupt saturation of torque.
[0086] Even better, during the calculation process in steps 304-307, when the vehicle already has a clear direction of movement and the operator applies a force in the opposite direction but has not yet exceeded the corresponding dead zone threshold, the system attenuates the first assist component according to the proportion of the reverse force to the current reverse dead zone threshold. This allows the operator to feel the reduction in the original direction assist before officially crossing the reverse dead zone, reducing the jerky feeling of sudden braking or sudden change of direction and improving the smoothness of handling.
[0087] In a preferred embodiment, the steps for slope adaptation adjustment include:
[0088] Step 311: Obtain information on the vehicle's bumpy condition;
[0089] Step 312: When it is determined that the vehicle is in a rough driving state based on the bumpy state information, limit or reduce the gain of the slope adaptation adjustment.
[0090] Step 313: When it is determined from the bumpy state information that the vehicle body is not in a rough driving state and the pitch attitude information exceeds the preset slope dead zone, the slope gain is calculated based on the amplitude of the pitch attitude information exceeding the slope dead zone, and the basic drive requirements are adjusted using the slope gain.
[0091] In steps 311-313 above, the vehicle's pitch angle is affected by vibration on bumpy roads. Direct slope compensation could lead to erroneous gain and sudden torque changes. Therefore, this solution actively limits or reduces the slope compensation gain when rough driving conditions are detected, avoiding erroneous torque output caused by vibration contamination of pitch angle measurements and improving control robustness. When the vehicle is stable and the pitch angle exceeds the slope dead zone, the slope gain is calculated based on the excess amplitude, achieving a safe compensation effect of increased force uphill and limited force downhill.
[0092] In one example, the determination of the bumpy state in steps 311-312 employs a hysteresis de-jitter strategy with fast entry and slow exit: the bumpy state is entered after 5 consecutive frames of detecting a bump indicator, and exits only after 20 consecutive frames of not detecting a bump indicator. This hysteresis strategy effectively avoids frequent switching of the bumpy state when there are slight road surface fluctuations, ensuring the stability of the slope compensation control.
[0093] In one example, the ramp gain in step 313 is set with an upper limit (e.g., 1.5 times) to prevent excessive compensation from causing the vehicle to lose control. At the same time, as the vehicle speed exceeds the preset decay start speed (e.g., 1.0 m / s), the ramp gain gradually decays to a smaller value at high speeds (e.g., 1.0 times) to avoid overcompensation affecting handling stability at high speeds.
[0094] In another example, step 313, in addition to providing slope compensation, can also superimpose a speed damping torque opposite to the direction of the rollover. This damping torque increases linearly with the overspeed after the actual wheel speed exceeds the preset anti-rollover trigger speed (e.g., 0.05 m / s) until it reaches the maximum damping torque limit. This speed damping constitutes an anti-rollover "speed safety net" that does not rely on the operator continuously increasing the force, providing additional safety protection when starting on a slope.
[0095] In a preferred embodiment, the steps for adjusting cornering safety include:
[0096] Step 321: When the absolute value of the yaw attitude information exceeds the preset cornering threshold, lock the current value or the previous cycle value of the basic drive requirement before entering the cornering state as the torque upper limit.
[0097] Step 322: During the turning state, the basic drive demand after slope adaptation adjustment is limited to no more than the upper limit of torque, and the drive demand is further attenuated according to the absolute value of the yaw attitude information, wherein the larger the absolute value of the yaw attitude information, the greater the attenuation.
[0098] Step 323: When the absolute value of the yaw attitude information is lower than the preset corner exit threshold, release the turning state and cancel the torque limit lock.
[0099] Steps 321-323 above primarily achieve active safety control during cornering through cornering torque locking and dynamic attenuation. Specifically, unlike existing technologies that simply reduce torque, this solution first locks the pre-corner drive demand value as the torque upper limit at the moment of cornering, ensuring that the drive torque during cornering does not exceed the pre-corner level, fundamentally preventing the risk of forward lurching caused by continued increase in assist after cornering. Based on this, the drive demand is then dynamically attenuated according to the yaw rate; the sharper the corner, the greater the attenuation, achieving dual safety protection of "locking the upper limit first, then attenuating with the corner." The hysteresis design of the cornering threshold and exit threshold avoids the vehicle frequently entering and exiting the cornering control mode near critical cornering states, improving control stability.
[0100] In one example, the entry threshold in steps 321-323 above can be set to 0.2 rad / s, and the exit threshold can be set to 0.1 rad / s, forming a hysteresis range to avoid frequent switching. The turning state begins when the absolute value of the yaw rate gradually increases from 0.1 rad / s to above 0.2 rad / s; after entering the turning state, it must drop below 0.1 rad / s before exiting, preventing repeated jumps near the critical value.
[0101] In one example, the attenuation ratio during cornering in step 322 can be gradually reduced from 1.0 to a preset minimum ratio (e.g., 0.5) based on the current yaw rate. This attenuation ratio is then processed by a first-order low-pass filter before being multiplied by the drive requirements. The low-pass filter smooths the rate of change of the attenuation coefficient, ensuring a smooth change in torque output during cornering. This avoids drive torque jitter caused by fluctuations in yaw rate, further improving handling quality during cornering.
[0102] Step 4, Soft Surface Adjustment: When in the gripping state and the terrain mode information is in the second mode, the target motion speed and feedforward drive component are determined based on the human-machine interaction force information. The target motion speed is independently controlled by closed-loop following based on the motion state information, generating closed-loop adjustment components for each wheel. The initial torque of each wheel is obtained based on the feedforward drive component and the closed-loop adjustment component. Then, the first anti-slip intervention is performed based on the rate of change of the motion state of each drive wheel, and the second anti-slip intervention is performed based on the difference in motion state between the left and right drive wheels. The activation condition of the second anti-slip intervention is constrained by the yaw attitude information. The initial torque is corrected based on the results of the first and second anti-slip interventions, and the second torque command is generated.
[0103] Unlike torque control used on hard surfaces, torque control alone on soft surfaces, due to their low coefficient of friction, can easily lead to wheel spin and sinking. This solution maps the operator's applied force to the target speed and uses closed-loop control to ensure the actual wheel speed follows the target speed, allowing the vehicle to move at a stable speed on low-friction surfaces. Furthermore, a first anti-slip intervention (based on single wheel angular acceleration) identifies sudden wheel spin, while a second anti-slip intervention (based on the difference in speed between the left and right wheels) identifies persistent imbalance. These two anti-slip criteria complement each other, and the second is gated by yaw attitude to avoid misjudging normal turns. This forms a low-cost anti-slip system that requires no additional speed sensors, solving the problems of difficult starting on soft surfaces such as sand and the tendency to get stuck during continuous driving.
[0104] In a preferred embodiment, the terrain pattern information in step 1 can be selected by the user or automatically identified by the controller. The automatic identification can be based on one or more of the following characteristics: wheel speed fluctuation (such as the standard deviation or peak change rate of the wheel speed signal), wheel angular acceleration (such as an abnormal increase in the angular acceleration of a single wheel), left and right wheel speed difference (such as a continuous significant difference between the left and right wheels), attitude vibration (such as the fluctuation amplitude of the vehicle body pitch or yaw rate), or drive load (such as the variation range of motor current).
[0105] In a preferred embodiment, before determining the target motion speed and feedforward drive component based on the human-machine interaction force information, the method further includes: step 401, determining the direction of the human-machine interaction force information;
[0106] Step 402: If the direction of the human-machine interaction force in the current cycle is opposite to the direction of the historical filtered force, enter the reversal confirmation window. During the reversal confirmation window, the historical filtered force is attenuated with the goal of zero and the original sign remains unchanged. Step 403: Only after the human-machine interaction force in the opposite direction has reached the preset number of confirmation frames is it allowed to establish the target motion speed in the opposite direction.
[0107] Steps 401-403 described above, through continuous commutation confirmation and a sign attenuation mechanism, avoid instantaneous reverse false triggers caused by bumps or sensor noise. Specifically, when an opposing force is detected, the system does not immediately respond to commutation but enters a confirmation window period. During this period, the original force direction remains unchanged, but its amplitude gradually attenuates towards zero, allowing the vehicle to decelerate smoothly. Only after the opposing force has persisted for a preset number of frames is the commutation truly completed. This mechanism filters out instantaneous reverse interference signals generated by bumpy road surfaces and achieves a smooth transition in direction switching, avoiding vehicle jerking and motor shock caused by frequent commutations.
[0108] In one example, the number of commutation confirmation frames in steps 402 and 403 above can be set to 3 frames (corresponding to 30ms, calculated based on a 10ms control cycle). Moreover, during the confirmation window, the historical filtering force approaches zero at a rate of 50% decay per frame while maintaining its original sign, so that the driving force fades smoothly rather than abruptly, significantly improving the smoothness of the commutation process.
[0109] In another example, once the reversal is confirmed, the filter state of the human-machine interaction force, the integral term of the left and right wheel speed closed loop, and the ramp state of the target speed must be cleared simultaneously to avoid residual torque impact from the historical state in the new direction.
[0110] In a preferred embodiment, after obtaining the initial torque of each wheel based on the feedforward drive component and the closed-loop adjustment component, the method further includes: step 411, initially setting the upper limit of the allowable torque of each wheel to the first upper limit of torque;
[0111] Step 412: When the rotational speed of any drive wheel is lower than the preset escape speed threshold, the allowable torque limit of the drive wheel is temporarily switched to a second torque limit that is greater than the first torque limit.
[0112] Step 413: When the speed of the drive wheel exceeds the speed threshold for getting out of trouble, restore the maximum allowable torque of the drive wheel to the first maximum torque.
[0113] Step 414, correcting the initial torque specifically includes: limiting the initial torque to no higher than the current allowable torque limit of the corresponding drive wheel.
[0114] Steps 411-414 above balance the seemingly contradictory needs of starting and escaping from sandy terrain and preventing getting stuck while driving continuously. Specifically, when the vehicle is stationary or at extremely low speeds, the system allows the drive wheels to briefly use a higher maximum torque to help the vehicle overcome static friction or escape from soft surfaces. Once the wheel speed exceeds the escaping threshold, indicating that the vehicle has successfully started or escaped, the system immediately switches the torque limit to a lower maximum to prevent the wheels from continuously spinning and sinking due to high torque. This solution ensures escaping capability during startup while avoiding the "digging" effect caused by continuous high torque on soft surfaces.
[0115] In one example, the first torque limit in step 411 above can be 60% of the rated torque, and the second torque limit can be 100% of the rated torque. The traction speed threshold can be set to 50 rpm (approximately 0.3 m / s corresponds to a wheel speed). That is, when the wheel speed is below 50 rpm, it is allowed to output the rated torque for a short time, and after exceeding 50 rpm, it automatically recovers to the continuous torque limit of 60%, effectively preventing the wheel from spinning and digging sand after the wheel speed increases.
[0116] In a preferred embodiment, the step of performing the first anti-slip intervention based on the rate of change of motion state of each drive wheel includes: step 421, determining the angular acceleration of the drive wheel based on the difference between the current cycle speed and the previous cycle speed of each drive wheel;
[0117] Step 422: Filter the angular acceleration;
[0118] Step 423: When the filtered angular acceleration of any drive wheel continuously exceeds a preset acceleration threshold in the direction of the target velocity, the drive wheel is determined to be slipping.
[0119] Step 424: When slippage occurs, reduce the closed-loop control integral term of the drive wheel and reduce the output torque of the drive wheel by a preset penalty ratio during the slippage period.
[0120] Steps 421-424 above utilize the existing motor speed feedback in the drive system to achieve sudden slippage detection without additional sensors. Specifically, the derivative of the motor speed (i.e., angular acceleration) directly reflects the rate of change of wheel speed. When a wheel loses traction and spins freely, its speed will increase sharply in a very short time, manifested as an abnormal increase in angular acceleration. Therefore, this solution avoids false triggering caused by noise spikes by filtering the angular acceleration and continuously judging exceeding thresholds. Once slippage is detected, the system immediately reduces the integral term of that wheel and continuously reduces the output torque, allowing the wheel to quickly regain traction. The torque automatically recovers after the slippage is eliminated. This solution can achieve single-wheel anti-slip without adding an additional vehicle speed sensor, effectively solving the problem of sudden single-wheel spinning on soft road surfaces while maintaining low cost.
[0121] In one example, the angular acceleration filtering in steps 421-423 uses an exponential moving average (EMA) filter with a filter coefficient of 0.3, balancing response speed and noise suppression. The acceleration threshold can be set to 2000 rpm / s (approximately equivalent to 0.2g of wheel surface acceleration). The number of consecutive frames exceeding the threshold can be set to 3 frames to avoid false triggering due to noise in a single frame.
[0122] In one example, in step 424, when a wheel enters a slipping state, the penalty ratio can be set to 0.5, which means reducing the current output torque of that wheel by 50%. During the slipping period, the accumulation of the integral term of the wheel's speed loop is paused to prevent the integral term from accumulating too much during the slipping period, which would cause a shock recovery of the output torque after the slipping ends.
[0123] In a preferred embodiment, the step of performing a second anti-slip intervention based on the difference in motion states between the left and right drive wheels includes:
[0124] Step 431: When the absolute value of the yaw attitude information is lower than the preset turning threshold, calculate the speed difference between the left drive wheel and the right drive wheel.
[0125] Step 432: When the absolute value of the speed difference exceeds the preset differential threshold, the drive wheel with the higher speed is identified as the slipping wheel;
[0126] Step 433: Determine the penalty coefficient based on the magnitude of the speed difference exceeding the differential threshold, and use the penalty coefficient to reduce the output torque of the slipping wheel and the integral term of the closed-loop control;
[0127] Step 434: When the absolute value of the yaw attitude information reaches or exceeds the turning threshold, skip the judgment of the speed difference in this cycle.
[0128] Steps 431-434 above, through the left-right wheel speed difference criterion of yaw attitude gating, accurately distinguish between normal turning and slippage. When a vehicle is actually turning, the outer wheel will inevitably turn faster than the inner wheel; this is a normal physical speed difference and should not be misjudged as slippage. This solution uses yaw rate to determine whether the vehicle is turning. Specifically, when driving straight, if the speed difference between the left and right wheels exceeds a threshold, it indicates that one side is slipping and should be penalized; when turning, this criterion is paused to avoid false triggering. This solution complements the angular acceleration criterion in the first anti-slip intervention. The first criterion focuses on sudden spinning of a single wheel, while the second criterion focuses on continuous imbalance between the left and right wheels. The two are integrated to form a complete dual-drive anti-slip system that does not require a ground speed sensor.
[0129] In one example, the turning threshold in steps 431-433 can be set to 0.15 rad / s. The differential threshold can be set to the higher-speed wheel exceeding the lower-speed wheel by 20%. The penalty coefficient can increase linearly with the deviation amplitude, with a penalty coefficient of 0.2 when the deviation is 20% and a penalty coefficient of 0.8 when the deviation is 50%, achieving the effect that the larger the differential, the stronger the penalty.
[0130] In one example, the angular acceleration anti-slip and electronic limited-slip differential mentioned in steps 432 and 433, respectively, can operate simultaneously and independently. The angular acceleration anti-slip is used to detect sudden increases in the speed of a single wheel and remains active in any driving state (including when turning); the electronic limited-slip differential is used to detect continuous imbalances in the speed of the left and right wheels and is activated only when driving straight and suspended when turning. The two have clear divisions of labor and do not interfere with each other, together forming a dual-drive anti-slip control system that does not require a ground speed sensor.
[0131] Step 5, Torque Control: Output the first torque command or the second torque command to the left drive wheel and the right drive wheel, respectively.
[0132] Step 5 of this solution sends independent torque commands to the left and right drive wheels, ensuring balanced power to both wheels when driving straight, differential speed coordination between the left and right wheels when turning, and individual torque limitation on the slipping wheel when one wheel slips. Compared to a single-motor drive system, this independent left and right drive method not only provides stronger power and better terrain adaptability, but more importantly, it provides independent execution channels for the first anti-skid intervention (single-wheel torque reduction) and the second anti-skid intervention (faster wheel torque reduction), making it a key execution link for the entire anti-skid control strategy to be implemented effectively.
[0133] In one example, before outputting torque commands to the left and right wheels, a maximum torque limit for the drive direction is constructed based on the current wheel speed: maximum torque output is allowed below the knee point speed; between the knee point speed and the zero torque speed, the maximum torque limit decreases linearly; and after reaching the zero torque speed, the maximum drive torque limit drops to zero. This external characteristic envelope of the motor ensures that the motor can deliver maximum torque for starting and getting out of trouble in the low-speed range, while automatically limiting output in the high-speed range to prevent overspeeding.
[0134] In another example, the output torque is also subject to dynamic ramp limiting: the ramp step size varies with vehicle speed, with a smaller step size (e.g., 0.5 Nm / ms) used at low speeds to achieve a smooth start, and a larger step size (e.g., 2.0 Nm / ms) used at high speeds to improve high-speed following performance; the ramp step size is fixed at a larger value (e.g., 3.0 Nm / ms) so that residual assist can be quickly withdrawn when the operator reduces the applied force, avoiding a dragging feeling.
[0135] Even better, when confirming mode switching (switching between hard and soft road surfaces) or direction switching, the corresponding controller's filter state, integral term, and ramp state must be reset to avoid sudden torque changes caused by historical states in the new control mode. This reset mechanism ensures torque continuity during mode and direction switching, eliminating the impact at the moment of switching.
[0136] The above embodiments are merely illustrative of the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent changes or modifications made based on the essence of the content of the present invention should be covered within the scope of protection of the present invention.
Claims
1. A control method for a two-wheeled electric bicycle, characterized in that, Includes the following steps: The system acquires information on the human-machine interaction force and grip state applied by the operator, the motion state information of the left and right drive wheels, the vehicle pitch attitude information and the vehicle yaw attitude information, and the terrain pattern information used to characterize the road surface adhesion conditions. The terrain pattern information includes at least a first mode and a second mode corresponding to hard road surface and soft road surface, respectively. Based on the gripping status information, it is determined whether the operator is in a gripping state. If the operator is not in a gripping state, the output of drive torque to the left and right drive wheels is terminated. When the vehicle is in a gripping state and the terrain mode information is the first mode, a first assist component that changes with vehicle speed is determined based on the motion state information, and a second assist component that changes with applied force is determined based on the human-machine interaction force information. A basic drive requirement is generated based on the first assist component and the second assist component. Then, based on the pitch attitude information and the yaw attitude information, the basic drive requirement is adjusted for slope adaptability and cornering safety, respectively, to generate a first torque command. When the machine is in a gripping state and the terrain mode information is the second mode, the target motion speed and feedforward drive component are determined according to the human-machine interaction force information. The target motion speed is independently controlled by closed-loop following according to the motion state information, and closed-loop adjustment components of each wheel are generated. The initial torque of each wheel is obtained according to the feedforward drive component and the closed-loop adjustment component. Then, a first anti-slip intervention is performed based on the rate of change of the motion state of each drive wheel, and a second anti-slip intervention is performed according to the difference in motion state between the left and right drive wheels. The activation condition of the second anti-slip intervention is constrained by the yaw attitude information. The initial torque is corrected according to the results of the first and second anti-slip interventions, and a second torque command is generated. The first torque command or the second torque command is output to the left drive wheel and the right drive wheel, respectively.
2. The control method for a two-wheeled electric bicycle according to claim 1, characterized in that, The step of terminating the output of drive torque to the left and right drive wheels includes: Execute braking action until the absolute value of the driving speed determined based on the motion state information is lower than the preset stopping speed; After coming to a stop, the absolute value of the pitch attitude information is used to determine: if the absolute value exceeds a first threshold, the braking lock state is maintained; if the absolute value is lower than a second threshold, the braking is released and the vehicle is allowed to move freely; if the absolute value is between the second threshold and the first threshold, the current state is maintained.
3. The control method for a two-wheeled electric bicycle according to claim 1, characterized in that, Before determining the first assist component that varies with vehicle speed based on the motion state information, the method further includes: Determine the current vehicle speed based on the motion state information; Based on the current vehicle speed, force dead zone thresholds corresponding to the operator's pushing and pulling forces are determined respectively. The force dead zone threshold when the current vehicle speed is zero or lower than the preset low speed threshold is greater than the force dead zone threshold when the current vehicle speed is higher than the preset travel speed threshold. When the amplitude of the human-machine interaction force information does not exceed the corresponding dead zone threshold of the interaction force, it is determined that there is no effective force application intention.
4. The control method for a two-wheeled electric bicycle according to claim 3, characterized in that, The step of generating basic driving requirements based on the first assist component and the second assist component includes: Determine the current vehicle speed based on the motion state information; The current vehicle speed is converted into the first assist component according to a preset segmented mapping curve. The segmented mapping curve includes at least: a zero assist section when the vehicle speed is below a first threshold, an increasing section when the vehicle speed is between the first threshold and the second threshold, a maintenance section when the vehicle speed is between the second threshold and the third threshold, and a decreasing or limiting section after the vehicle speed exceeds the third threshold. The human-machine interaction force information is normalized after deducting the corresponding dead zone threshold, and the normalized value is converted into the second assist component through a preset nonlinear mapping. The first assist component and the second assist component are added together to obtain the basic drive requirement.
5. The control method for a two-wheeled electric bicycle according to claim 1, characterized in that, The steps for slope adaptive adjustment include: Obtain the bump status information of the vehicle body; When the vehicle is determined to be in a rough driving state based on the bumpy state information, the gain of the slope adaptive adjustment is limited or reduced. When it is determined from the bumpy state information that the vehicle body is not in a rough driving state and the pitch attitude information exceeds the preset slope dead zone, the slope gain is calculated based on the magnitude of the pitch attitude information exceeding the slope dead zone, and the basic drive requirements are adjusted using the slope gain.
6. The control method for a two-wheeled electric bicycle according to claim 1, characterized in that, The steps for adjusting cornering safety include: When the absolute value of the yaw attitude information exceeds the preset cornering threshold, the current value or the previous cycle value of the basic drive requirement before entering the turning state is locked as the torque upper limit. During the turning state, the basic drive demand after the slope adaptive adjustment is limited to no more than the torque limit, and the drive demand is further attenuated according to the absolute value of the yaw attitude information, wherein the larger the absolute value of the yaw attitude information, the greater the attenuation. When the absolute value of the yaw attitude information is lower than the preset corner exit threshold, the turning state is released and the torque limit lock is canceled.
7. The control method for a two-wheeled electric bicycle according to claim 1, characterized in that, Before determining the target motion velocity and feedforward drive component based on the human-machine interaction force information, the method further includes: The direction of the human-machine interaction force information is determined; If the direction of the human-machine interaction force in the current cycle is opposite to the direction of the historical filtered force, the system enters the reversal confirmation window. During the reversal confirmation window, the historical filtered force is attenuated with the goal of zero while maintaining its original sign. The target motion speed in the opposite direction is only allowed to be established after the human-machine interaction force in the opposite direction has been continuously maintained for a preset number of confirmation frames.
8. The control method for a two-wheeled electric bicycle according to claim 1, characterized in that, After obtaining the initial torque of each wheel based on the feedforward drive component and the closed-loop adjustment component, the method further includes: The maximum allowable torque for each wheel is initially set to the first maximum torque. When the speed of any drive wheel is lower than the preset speed threshold for getting out of trouble, the allowable torque limit of that drive wheel is temporarily switched to a second torque limit that is greater than the first torque limit. When the rotational speed of the drive wheel exceeds the escape speed threshold, the allowable torque limit of the drive wheel is restored to the first torque limit. The modification of the initial torque specifically includes: limiting the initial torque to no higher than the current allowable torque limit of the corresponding drive wheel.
9. The control method for a two-wheeled electric bicycle according to claim 1, characterized in that, The step of performing the first anti-slip intervention based on the rate of change of the motion state of each drive wheel includes: The angular acceleration of each drive wheel is determined based on the difference between the current cycle speed and the previous cycle speed. The angular acceleration is filtered. When the filtered angular acceleration of any drive wheel continuously exceeds a preset acceleration threshold in the direction of the target motion velocity, the drive wheel is determined to be slipping. When the slippage state is entered, the closed-loop control integral term of the drive wheel is reduced, and the output torque of the drive wheel is reduced by a preset penalty ratio during the slippage period.
10. The control method for a two-wheeled electric bicycle according to claim 1, characterized in that, The step of performing the second anti-slip intervention based on the difference in motion state between the left and right drive wheels includes: When the absolute value of the yaw attitude information is lower than the preset turning threshold, the speed difference between the left drive wheel and the right drive wheel is calculated. When the absolute value of the speed difference exceeds a preset differential threshold, the drive wheel with the higher speed is identified as the slipping wheel; a penalty coefficient is determined based on the magnitude by which the speed difference exceeds the differential threshold, and the penalty coefficient is used to reduce the output torque of the slipping wheel and the closed-loop control integral term; When the absolute value of the yaw attitude information reaches or exceeds the turning threshold, the judgment of the speed difference in this cycle is skipped.
