Commercial vehicle cruise control method and system that takes into account both physiological strain and energy saving

Abstract

The present invention discloses a cruise control method and system for commercial vehicles that takes into account both physiological strain and energy conservation. The method includes the following steps: constructing a human physiological strain index, constructing a vehicle fuel consumption and energy conservation index, establishing an optimal cruise speed planning model, and comprehensively considering the vehicle fuel consumption and energy conservation index and travel time to form an overall reward function; introducing a reinforcement learning algorithm to solve and obtain the optimal cruise speed within the predicted time; constructing a longitudinal and vertical coupled dynamic model of the vehicle-seat-human system to calculate the human physiological strain index at the optimal cruise speed; encoding the vehicle state constraint and the optimal cruise speed sequence tracking constraint as control constraints, and forming a longitudinal model predictive control strategy based on the human physiological strain index, the vehicle fuel consumption and energy conservation index, and the tracking deviation; and adding a pitch control mechanism to realize vertical skyhook hybrid control and regulate the vehicle suspension. The present invention can improve vehicle cruise performance and achieve coordinated optimization of human physiological strain and vehicle energy conservation.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle cruise control, and in particular to a commercial vehicle cruise control method and system that takes into account both physiological strain and energy saving. Background Art

[0002] Commercial vehicles are primarily used for long-distance logistics transport. To alleviate the workload of drivers and achieve energy conservation and consumption reduction, predictive adaptive cruise control technology for commercial vehicles has become a research hotspot. Predictive adaptive cruise control is a key driver assistance method for energy-saving control in commercial vehicles. It uses on-board navigation or a cloud platform to obtain road information, including slope, curvature, and road conditions. Based on the vehicle's dynamic characteristics and fuel consumption model, it plans a fuel-efficient speed sequence, thereby assisted autonomous driving and minimizing fuel consumption.

[0003] While conventional predictive adaptive cruise control technology reduces driver workload, it lacks consideration for factors affecting comfort, such as vibration excitation, leading to the continued problem of physiological strain. Currently, the main method widely used to reduce vibration excitation and improve the driver's working environment is passive methods such as optimizing the cab's environmental layout, such as adjusting seat position and improving seat structure. These passive methods essentially minimize the impact of vibration excitation on the human body through various means after it occurs, but they cannot actively and directly eliminate the excitation source.

[0004] In recent years, a new approach has emerged: optimizing the behavior of long-distance commercial vehicles through motion planning and control, thereby reducing or eliminating excitation sources and improving the driver's working environment. For example, Saruch et al. used the rate of change of acceleration, or jerkiness, to analyze vehicle vibration in real time and optimize speed to ensure driving comfort and safety. However, based on research on human comfort mechanisms, the continuous stimulation of low-frequency acceleration is more likely to induce physiological strain. Therefore, the above-mentioned approach, which only considers jerkiness, is incomplete in reducing physiological strain and improving vibration comfort. Other path planning-based approaches, such as Chen et al., use path planning to avoid dense traffic flow, thereby reducing vehicle speed changes and the resulting pitch vibration. However, these approaches are inherently one-dimensional and static, attempting to fix the driving path or speed sequence at the planning start point. Without a real-time multidimensional feedback control strategy, they cannot dynamically adapt to changing traffic conditions, and thus pose certain risks during driving. Summary of the Invention

[0005] In order to overcome the defects and shortcomings of the existing technology, the present invention provides a commercial vehicle cruise control method and system that takes into account both physiological strain and energy saving. While improving the vehicle's fuel economy, the present invention dynamically adapts to traffic conditions and actively reduces vibration excitation to address the deficiency that traditional cruise functions do not pay enough attention to human physiological strain, further improve vehicle cruise performance, and achieve optimization of both human physiological strain and vehicle energy saving.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] The present invention provides a cruise control method for commercial vehicles that takes into account both physiological strain and energy saving, comprising the following steps:

[0008] Construct human physiological strain indicators, construct vehicle fuel consumption and energy-saving indicators, establish an optimal cruising speed planning model, and construct an overall reward function based on vehicle fuel consumption and energy-saving indicators and travel time;

[0009] The solution is based on the reinforcement learning algorithm. The state, acceleration, and reward values ​​are stored in the predicted experience playback area. During the training process, samples are taken from the predicted experience playback area to update the policy network and value network of the reinforcement learning algorithm in the prediction time domain, and solve the optimal cruising speed within the prediction time.

[0010] A longitudinal and vertical coupled dynamic model of the vehicle-seat-human system was constructed. Acceleration was calculated based on the optimal cruising speed. The acceleration and road excitation were input into the longitudinal and vertical coupled dynamic model of the vehicle-seat-human system to obtain vertical vibration acceleration and pitch angular acceleration. The physiological strain index of the human body at the optimal cruising speed was calculated.

[0011] The vehicle state constraints and optimal cruising speed sequence tracking constraints are encoded as control constraints, and a longitudinal model predictive control strategy is formed based on human physiological strain indicators, vehicle fuel consumption and energy saving indicators, and tracking deviation.

[0012] The additional pitch control mechanism realizes vertical skylight hybrid control and regulates the vehicle suspension.

[0013] As a preferred technical solution, the human physiological strain index is constructed, specifically including:

[0014] The vibration excitation is transformed into frequency domain and weighted. In the frequency domain transformation process, the vertical vibration acceleration is transformed into , pitch vibration acceleration Transformed into vertical vibration acceleration spectrum function , pitch vibration acceleration spectrum function In the weighted processing process, corresponding weighting functions are provided for vibrations in different axes. The weighted vibration acceleration spectrum function is:

[0015] ;

[0016] ;

[0017] in, 、 are the vertical weighted vibration acceleration spectrum function and the pitch weighted vibration acceleration spectrum function, is the vertical vibration weighting function, is the pitch vibration weighting function, 、 is the corresponding weight coefficient;

[0018] Based on the inverse Fourier transform 、 Transformed into vertical weighted vibration acceleration , pitch weighted vibration acceleration ;

[0019] The quantitative calculation formula for the impact of vertical and pitch vibration excitation on the human body is constructed as follows:

[0020] ;

[0021] ;

[0022] in, represents the vertical weighted vibration impact value, represents the pitch weighted vibration impact value, Indicates duration;

[0023] The human body physiological strain index is expressed as:

[0024] ;

[0025] in, Indicates the physiological strain index of the human body. 、 、 、 is the model coefficient.

[0026] As a preferred technical solution, a vehicle fuel consumption and energy saving index is constructed, which is specifically expressed as follows:

[0027] ;

[0028] ;

[0029] in, Indicates the vehicle's fuel consumption and energy saving index. Indicates the engine output power, 、 、 are coefficients to be fitted, is the engine speed, is the vehicle speed, is the frontal area of ​​the vehicle, is the drag coefficient, is the vehicle gravity, is the road slope, is the vehicle weight, is the longitudinal acceleration, is the rolling resistance coefficient, The vehicle transmission efficiency.

[0030] As a preferred technical solution, an optimal cruising speed planning model is established, specifically including:

[0031] The iterative process of the predictive cruise speed planning strategy is expressed in discrete form as follows:

[0032] ;

[0033] in, For the The vehicle position during step planning, For the The vehicle speed during step planning, For the The vehicle position during step planning, For the The vehicle speed during step planning, For the Vehicle acceleration during step planning, For the The air resistance acceleration caused by air resistance during step planning, is the acceleration due to gravity, is the friction coefficient, For the The road slope corresponding to the vehicle's position during step planning, The planning step duration.

[0034] As a preferred technical solution, an overall reward function is constructed based on vehicle fuel consumption and energy-saving indicators and travel time, which is expressed as:

[0035] ;

[0036] in, and is the adjustment coefficient, Indicates the vehicle's fuel consumption and energy-saving index.

[0037] As a preferred technical solution, a solution is performed based on a reinforcement learning algorithm. The specific steps include:

[0038] Add a prediction experience playback area for the TD3 reinforcement learning algorithm And store the state, acceleration, and reward value in the predicted experience playback area, and form an array based on the state, acceleration, and reward value ,in, Indicates the The state of the step solution, is the corresponding acceleration, is the reward value, is the next state;

[0039] In the policy network and value network of the TD3 reinforcement learning algorithm, the policy network is used to receive current state data, combine and return features to output continuous actions, and the value network obtains sensor input and the action output of the policy network in the current state, and outputs the optimal value of the current state action;

[0040] In each planning step, the overall reward function is calculated, and the motion parameters for the next moment are calculated by combining the current state and the cruise speed planning model;

[0041] Based on the spline interpolation algorithm, intermediate points are inserted between adjacent planning steps to obtain the optimal cruising speed within the prediction time.

[0042] As a preferred technical solution, a longitudinal and vertical coupled dynamic model of the vehicle-seat-human system is constructed, which is specifically expressed as follows:

[0043] ;

[0044] in:

[0045] 、 、 、

[0046] 、 、 、

[0047] 、 、 、

[0048] ;

[0049] in, are the moments of inertia of each part, are the vertical vibration accelerations of each part, are the vertical displacements of each part, and are the vertical displacements of the front and rear suspensions of each part, and are the vertical velocities of the front and rear suspensions of each part, are the pitch displacements of each part, are the pitch angular accelerations of each part, For the quality of each part, is the height of the centroid of each part, and are the front and rear suspension spring stiffness of each part respectively, and are the damping coefficients of the front and rear suspension of each part, and are the distances from the center of mass of each part to the front and rear suspensions, Indicates the pitch angle of each part, subscript Corresponding to the vehicle frame, cab and seat-human body system respectively, and are the displacements of the front and rear suspension lower ends of the seat, and are the displacements of the front and rear suspension lower ends of the cab, and are the distances from the cab center of mass to the front and rear suspensions of the seat, and are the distances from the center of mass of the frame to the front and rear suspensions of the cab, respectively.

[0050] As a preferred technical solution, the vehicle state constraint is expressed as:

[0051] ;

[0052] in, and The minimum and maximum limits of vehicle acceleration are: and The minimum and maximum vehicle speed limits are: is the actual distance between the vehicle and the preceding vehicle, is the minimum safe distance between vehicles. is the time distance between the vehicle and the preceding vehicle;

[0053] Based on the tracking deviation between the actual vehicle speed and the planned optimal cruising speed Construct a tracking constraint, expressed as:

[0054] ;

[0055] in, is the vehicle speed error, is the acceleration error, is the acceleration error weight coefficient;

[0056] The objective function of the longitudinal model predictive control strategy is expressed as:

[0057] ;

[0058] in, 、 、 is the weight factor, Indicates the physiological strain index of the human body. Indicates the vehicle's fuel consumption and energy-saving index.

[0059] As a preferred technical solution, an additional pitch control mechanism is used to achieve vertical ceiling hybrid control, specifically including:

[0060] The vertical ceiling control strategy is designed in sections according to different working conditions, which can be expressed as:

[0061] ;

[0062] in, represents the damping coefficient, represents the vertical acceleration, represents the vertical acceleration threshold, represents the maximum damping coefficient, represents the minimum damping coefficient, Indicates the physiological strain index of the human body;

[0063] The additional pitch control strategy is:

[0064] ;

[0065] in, and are the front and rear seat suspension damping coefficients, is the damping torque distribution coefficient of the front and rear seat suspension, is the distance between the front and rear seat suspensions, is the vertical velocity of the front suspension of the seat-body system, is the vertical velocity of the rear suspension of the seat-body system, Indicates the pitch angular velocity.

[0066] The present invention also provides a commercial vehicle cruise control system that takes into account both physiological strain and energy conservation, and is used to implement the commercial vehicle cruise control method that takes into account both physiological strain and energy conservation, comprising: a human physiological strain index module, a vehicle fuel consumption and energy conservation index module, an optimal cruise speed planning module, a reinforcement learning solution module, a dynamic model module, a longitudinal model predictive control module, and a hybrid control module;

[0067] The human body physiological strain index construction module is used to construct the human body physiological strain index;

[0068] The vehicle fuel consumption and energy saving index construction module is used to construct the vehicle fuel consumption and energy saving index;

[0069] The optimal cruising speed planning module is used to construct an optimal cruising speed planning model and build an overall reward function based on vehicle fuel consumption and energy saving indicators and travel time;

[0070] The reinforcement learning solution module is used to optimize the reinforcement learning algorithm and solve the optimal cruising speed. The state, acceleration, and reward value are stored in the predicted experience playback area. During the training process, samples are taken from the predicted experience playback area to update the policy network and value network of the reinforcement learning algorithm in the prediction time domain, and solve the optimal cruising speed within the prediction time.

[0071] The dynamic model building module is used to build a longitudinal and vertical coupled dynamic model of the vehicle-seat-human body system, calculate acceleration based on the optimal cruising speed, input the acceleration and road excitation into the longitudinal and vertical coupled dynamic model of the vehicle-seat-human body system to obtain vertical vibration acceleration and pitch angular acceleration, and calculate the human physiological strain index at the optimal cruising speed;

[0072] The longitudinal model predictive control module is used to encode the vehicle state constraints and the optimal cruising speed sequence tracking constraints as control constraints, and form a longitudinal model predictive control strategy based on human physiological strain indicators, vehicle fuel consumption and energy saving indicators, and tracking deviation;

[0073] The hybrid control module is used to add a pitch control mechanism to realize vertical skyhook hybrid control and regulate the vehicle suspension.

[0074] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0075] (1) This paper explores the response characteristics of the human body to different axial vibrations and proposes a human physiological strain index that takes into account multi-axial vibration excitation and human perception sensitivity. At the same time, it uses road slope parameters to expand the traditional fuel consumption calculation method and proposes a more accurate vehicle fuel consumption and energy saving index, laying a target parameter foundation for constructing a dual-objective control method that covers human physiological strain and vehicle fuel consumption economy.

[0076] (2) Based on the advantages of predictive adaptive cruise control technology, the present invention integrates map information such as road slope and speed limit, proposes an energy-saving and efficient speed planning model, and combines the predicted experience replay mechanism to improve the reinforcement learning solution algorithm, which can significantly improve the speed planning time domain length and the comprehensive performance of vehicle cruise control, and provide the optimal cruise speed reference sequence for subsequent control.

[0077] (3) The present invention constructs a longitudinal and vertical coupling dynamic model of the vehicle-seat-human system by considering the coupling transfer characteristics of vibration excitation. Based on the constraint state and control process that can be automatically activated or gradually weakened, the vehicle can intelligently and adaptively adjust the control output based on the real-time traffic situation to achieve a comprehensive performance improvement that takes into account both physiological strain and energy saving. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 A flow chart of a commercial vehicle cruise control method that takes into account both physiological strain and energy saving according to the present invention;

[0079] Figure 2 This is a schematic diagram comparing the actual fuel consumption rate and the fitted value of the present invention;

[0080] Figure 3 This is a schematic diagram of the overall architecture of the reinforcement learning solution algorithm of the present invention;

[0081] Figure 4 Schematic diagram of the overall architecture of the longitudinal and vertical coupled dynamic model of the vehicle-seat-human body system of the present invention;

[0082] Figure 5 A schematic diagram of road elevation changes in a simulation verification scenario of the present invention;

[0083] Figure 6 Schematic diagram of the vehicle speed curve of the present invention;

[0084] Figure 7 This is a schematic diagram of the comparison of vehicle accelerations of the present invention;

[0085] Figure 8 This is a schematic diagram comparing the fuel consumption of the present invention;

[0086] Figure 9 Schematic diagram comparing the damping force of the front shock absorber of a vehicle before and after applying the control method of the present invention;

[0087] Figure 10 Schematic diagram comparing the damping force of the rear shock absorber of a vehicle before and after applying the control method of the present invention;

[0088] Figure 11 This is a schematic diagram of the improvement effect of the human physiological strain indicators of the present invention. DETAILED DESCRIPTION

[0089] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0090] Example 1

[0091] like Figure 1As shown, this embodiment provides a cruise control method for a commercial vehicle that takes into account both physiological strain and energy saving, including the following steps:

[0092] S1: Construct human physiological strain index;

[0093] During commercial vehicle cruising, vertical and pitch vibrations are the primary vibration excitations experienced by the human body. The coupling of these two vibrations is the root cause of physiological strain. Therefore, based on the coupling of vertical and pitch vibrations, as well as the human body's response to vibrations in different axial directions and at different frequencies, a human physiological strain index under forced vibration was constructed, laying the foundation for target parameters for the subsequent construction of control strategies.

[0094] Since the human body responds differently to vibrations of different axes and frequencies, in order to more accurately evaluate the impact of vibration on the human body, the vibration excitation is first transformed in the frequency domain and weighted. In the frequency domain transformation process, the vertical vibration acceleration is converted into , pitch vibration acceleration Transformed into vertical vibration acceleration spectrum function , pitch vibration acceleration spectrum function In the weighted processing process, corresponding weighting functions are provided for vibrations in different axes, among which the vertical vibration weighting function is , the pitch vibration weighting function is , the weighted vibration acceleration spectrum function is:

[0095] ;

[0096] ;

[0097] in, 、 are the vertical weighted vibration acceleration spectrum function and the pitch weighted vibration acceleration spectrum function, 、 is the corresponding weight coefficient, and then the inverse Fourier transform method is used to transform 、 Transformed into vertical weighted vibration acceleration , pitch weighted vibration acceleration ;

[0098] Taking into account the amplitude, frequency and time history of vibration acceleration, the quantitative calculation formula for the impact of vertical and pitch vibration excitation on the human body is constructed as follows:

[0099] ;

[0100] ;

[0101] in, represents the vertical weighted vibration impact value, represents the pitch weighted vibration impact value, Indicates duration;

[0102] Taking into account the physiological strain indicators of the human body The weighted vibration dose value and vibration duration Therefore, the comprehensive calculation formula of human physiological strain index is expressed as:

[0103] ;

[0104] in, 、 、 、 is the model coefficient;

[0105] S2: Construct vehicle fuel consumption and energy saving indicators;

[0106] The vehicle fuel consumption dataset serves as the data foundation for research on transient fuel consumption models, vehicle fuel consumption and energy-saving indicators, and subsequent optimal cruise speed planning. To this end, a targeted simulation scheme was designed, and vehicle trajectories, fuel consumption data, and road slopes were recorded. Subsequently, impurity data removal, data smoothing, and feature extraction were performed on the trajectory data. The fuel consumption data was then integrated to form a vehicle fuel consumption dataset, as shown in Table 1 below.

[0107] Table 1 Vehicle fuel consumption dataset format

[0108]

[0109] The traditional instantaneous fuel consumption index is constructed based on speed and acceleration. However, analysis has found that road slope also has a certain impact on vehicle fuel consumption. Therefore, the traditional fuel consumption calculation method is expanded by using road slope. That is, based on the multivariate regression method and the BIC criterion (Bayesian Information Criterion), the vehicle fuel consumption energy saving index is constructed by integrating speed, acceleration, engine speed and road slope. Considering the complexity of the calculation formula, terms with a degree of more than 3 are not considered, so the transient fuel consumption energy saving index is defined as follows: The basic calculation structure is:

[0110] ;

[0111] Engine output power Expressed as:

[0112] ;

[0113] in, is the engine speed, is the vehicle speed, is the frontal area of ​​the vehicle, is the drag coefficient, is the vehicle gravity, is the road slope, is the vehicle weight, is the longitudinal acceleration, is the rolling resistance coefficient, is the vehicle transmission efficiency (0.85 for commercial vehicles), 、 、 are coefficients to be fitted;

[0114] Use the vehicle fuel consumption data set to solve the coefficients to be fitted and obtain the final transient fuel consumption energy saving index The calculation formula is:

[0115] ;

[0116] Randomly extract some data to verify the accuracy of the calculation formula, such as Figure 2 As shown, the verification results are obtained, where R 2 is 0.97, which means that the actual fuel consumption is close to the calculated fuel consumption. The calculation formula can better estimate the actual instantaneous fuel consumption rate;

[0117] S3: Optimal cruising speed planning modeling;

[0118] The advantage of predictive cruise control is that it can predictably obtain and utilize road ahead information, and design a predictive reward function based on this information, thereby achieving long-term planning of economical cruise speed. Road ahead information is obtained from the vehicle's onboard geographic information system and processed into a function that relates the vehicle's position to the road slope angle and speed limit, namely:

[0119] ;

[0120] in, is the road slope sequence, Indicates the road slope For location The discrete function of For road speed limits, Indicates the road speed limit For location Discrete function, combined with the future road slope sequence and road speed limits , construct the observation state vector of the predictive cruise speed planning strategy for:

[0121] ;

[0122] On this basis, the iterative process of the predictive cruise speed planning strategy is expressed in discrete form, namely:

[0123] ;

[0124] in, For the The vehicle position during step planning, For the The vehicle speed during step planning, For the Vehicle acceleration during step planning, For the The air resistance acceleration caused by air resistance during step planning, is the acceleration due to gravity, is the friction coefficient, For the The road slope corresponding to the vehicle's position during step planning, The planning step duration.

[0125] In order to realize the economic cruising speed planning based on the road ahead information, a reasonable reward function must be designed to ensure economy and timeliness. Ensure fuel economy and timeliness based on travel time, so the overall reward function for:

[0126] ;

[0127] in, and is the adjustment coefficient.

[0128] Output of the predictive cruise speed planning strategy The constraints are:

[0129] ;

[0130] The above constraints mean that the economic cruising speed planned by the predictive cruising speed planning strategy must be within the reference speed A certain range Internal floating, no drastic speed fluctuation output;

[0131] S4: Solve based on reinforcement learning algorithm;

[0132] Although the TD3 reinforcement learning algorithm (Twin Delayed Deep Deterministic Policy Gradient Algorithm) can solve problems with continuous action space and high-dimensional state space, it can only plan the speed of the next cycle and cannot perform speed planning in the long term. Figure 3 As shown, add a prediction experience playback area for the TD3 algorithm , in the training process, the prediction solution based on the deterministic mathematical model is executed, and the solution is arrays Store in the prediction experience playback area, where Indicates the The state of the step solution (including position, speed and other information), is the corresponding acceleration, is the reward value, is the next state; at the same time, during the training process Sampling enables updates to the TD3 algorithm's policy network and value network within the prediction domain. In the TD3 reinforcement learning algorithm's Actor-Critic network structure, the Actor (policy network) receives current state data, combines and returns features to output continuous actions, while the Critic (value network) obtains sensor input and the Actor (policy network)'s action output in the current state, ultimately outputting the optimal value for the current state and action.

[0133] After the TD3 reinforcement learning algorithm is trained, it is used to solve the optimal cruising speed within the predicted duration. Specifically, in each step of the planning solution, the TD3 reinforcement learning algorithm uses the overall reward function To take into account both fuel consumption and timeliness, and at the same time calculate the motion parameters of the next moment according to the current state and the iterative expression in discrete form, that is, At the same time, the spline interpolation algorithm is used to insert intermediate points between adjacent planning steps to make the speed change more continuous and smooth, avoiding sudden changes.

[0134] S5: Longitudinal and vertical coupled dynamic modeling of vehicle-seat-human system;

[0135] Vertical vibration and pitch vibration are the main factors affecting human physiological strain. Based on this premise, a vertical and vertical coupling dynamic model of the vehicle-seat-human system is constructed, such as Figure 4 As shown in the figure, the vertical and longitudinal coupled dynamic model of the vehicle-seat-human system includes three parts, subscript They correspond to the vehicle frame, cab and seat-human body system respectively. The vertical excitation of the vehicle frame, cab, and seat-human body system all generate pitch and vertical displacement due to the vertical excitation of the road. Based on this, the longitudinal and vertical coupled dynamic model of the vehicle-seat-human body system is constructed as follows:

[0136] ;

[0137] in:

[0138] 、 、 、

[0139] 、 、 、

[0140] 、 、 、

[0141] ;

[0142] in, are the moments of inertia of each part, , Indicates the pitch angle of each part, and are the displacements of the front and rear suspension lower ends of the seat, and are the displacements of the front and rear suspension lower ends of the cab, are the vertical vibration accelerations of each part, are the vertical displacements of each part, and are the vertical displacements of the front and rear suspensions of each part, and are the vertical velocities of the front and rear suspensions of each part, are the pitch displacements of each part, are the pitch angular accelerations of each part, For the quality of each part, is the height of the centroid of each part, and are the front and rear suspension spring stiffness of each part respectively, and are the damping coefficients of the front and rear suspension of each part, and are the distances from the center of mass of each part to the front and rear suspensions, and are the distances from the cab center of mass to the front and rear suspensions of the seat, and are the distances from the center of mass of the frame to the front and rear suspensions of the cab, respectively. The specific parameter values ​​of the above parameters are shown in Table 2 below:

[0143] Table 2 Vehicle structure parameters

[0144]

[0145] After completing the construction of the longitudinal and vertical coupled dynamic model of the vehicle-seat-human system, the acceleration is calculated through the planned optimal cruising speed sequence , and the acceleration The vertical vibration acceleration of the seat-human body system can be output by inputting the vehicle-seat-human body system longitudinal and vertical coupling dynamic model with road excitation. Pitch angular acceleration Substitute these two values ​​into the human physiological strain index In the calculation process, , , we can get the specific stress that the human body endures when the vehicle tracks the planned optimal cruising speed. value, the The value is used as a target in the subsequent control part to improve vibration comfort and reduce human physiological strain;

[0146] S6: Constructing a longitudinal model predictive control strategy;

[0147] When tracking the planned optimal cruising speed sequence, the vehicle state constraints and the optimal cruising speed sequence tracking constraints are encoded as control constraints. At the same time, a longitudinal model predictive control strategy is formed by combining human physiological strain indicators, vehicle fuel consumption and energy saving indicators, and tracking deviations to achieve adaptive switching of control modes under different working conditions, while improving vehicle fuel economy and human comfort.

[0148] Vehicle state constraints usually include acceleration limits, speed limits, and safe vehicle distance constraints, namely:

[0149] ;

[0150] in, and The minimum and maximum limits of vehicle acceleration are: and The minimum and maximum vehicle speed limits are: is the actual distance between the vehicle and the preceding vehicle, is the minimum safe distance between vehicles. is the time distance between the vehicle and the preceding vehicle;

[0151] To ensure that the vehicle tracks the planned optimal cruise speed sequence, the tracking deviation between the actual speed and the planned optimal cruise speed is calculated. To construct a tracking constraint:

[0152] ;

[0153] in, is the vehicle speed error, is the acceleration error, is the acceleration error weight coefficient;

[0154] When constructing the objective function, the human physiological strain index, vehicle fuel consumption and energy saving index, and tracking deviation are integrated. The objective function expression of the longitudinal model predictive control strategy is:

[0155] ;

[0156] in, 、 、 is the weight factor, which is used to balance the weight relationship between the index values;

[0157] For the proposed longitudinal model predictive control strategy, the vehicle state constraints exist in two states: activated or inactive, while the tracking constraints are always in the activated state. By determining whether the vehicle state constraints are activated, the vehicle can implicitly switch between the two modes of tracking the planned optimal cruising speed sequence and maintaining a safe distance from the vehicle in front. Specifically, when there is no vehicle ahead or the distance is far beyond the minimum safe distance, the constraint amount tends to infinity, that is, the constraint tends to fail, and the control strategy naturally degenerates into the optimal cruising speed sequence tracking control subject only to the tracking constraint. The vehicle mainly focuses on tracking the planned optimal cruising speed sequence; when there is a vehicle ahead, in order to ensure driving safety, the vehicle distance constraint is strengthened, forcing the longitudinal model predictive control strategy to output an acceleration that satisfies the requirement of maintaining a safe distance. , that is, switching to a driving mode that takes into account both safe vehicle distance and tracking deviation.

[0158] S7: Construct a hybrid control strategy for vertical ceilings;

[0159] When the vehicle is tracking the planned optimal cruising speed sequence, it includes two operating conditions: constant speed driving and variable speed driving. In the constant speed driving process, the vertical vibration from the road is mainly suppressed. The front and rear suspensions do not interfere with each other and can be controlled completely independently. When changing speed, Substituting into the longitudinal and vertical coupled dynamic model of the vehicle-seat-human system, the pitch angular acceleration of the seat-human system can be obtained , that is, there is a pitch coupling between the front and rear seat suspensions, and it is difficult to achieve the best control effect by independently controlling each sub-suspension. Therefore, on the basis of the vertical skyhook control strategy, a pitch control strategy is added to form a vertical skyhook hybrid control strategy. First, a pitch angular acceleration threshold is set. , and Determines the activation state of the additional pitch control strategy. < , that is, when there is only vertical vibration, the front and rear seat suspensions are regulated by independent vertical skyhook control strategies; and when ≥ When pitch-coupled vibration exists, the pitch control strategy is activated to compensate for the damping torque distribution deviation caused by the joint control of the front and rear seat suspensions.

[0160] For the vertical ceiling control strategy, segmented design is performed according to different working conditions. When , it means that there is a large transient impact. In order to quickly dissipate the impact energy and suppress the violent vibration of the tire, the maximum damping coefficient is used. For other working conditions, the human physiological strain index is used Make the damping coefficient at the maximum value and minimum value Adaptive smooth adjustment between them, the control strategy expression is:

[0161] ;

[0162] Among them, human physiological strain index Based on vertical vibration acceleration Pitch angular acceleration calculate, and The output of the vehicle-seat-human system longitudinal and vertical coupling dynamic model is substituted into In the calculation process, , , we can get the specific stress that the human body endures when the vehicle tracks the planned optimal cruising speed. value.

[0163] when ≥ When the target damping torque is opposite to the pitch angular velocity, it is necessary to provide , so that the seat-human body system can return to stability as soon as possible. Since the target damping torque is determined by the controllable damping force of the front and rear seat suspension shock absorbers 、 Coupling occurs, and considering the difference in front and rear distribution of damping torque, the additional pitch control strategy is:

[0164] ;

[0165] In the formula and are the front and rear seat suspension damping coefficients, is the damping torque distribution coefficient of the front and rear seat suspension, is the distance between the front and rear seat suspensions, Indicates the pitch angular velocity.

[0166] This embodiment selects a highway in a certain city as a simulation verification scenario, such as Figure 5 As shown in , it is the road elevation change. Figure 6 As shown in the figure, the planned speed curve and control tracking effect are obtained. Among them, the dotted line is the energy-saving cruising speed planned only for energy saving. It can be observed that it is almost all near the minimum speed limit, resulting in a significant reduction in driving efficiency. Therefore, in order to achieve the best comprehensive performance of the vehicle, the balanced cruising speed that takes into account multiple indicators is used as the planned optimal cruising speed. The dotted line represents the control tracking speed. It can be observed that the designed longitudinal model predictive control strategy can effectively track the planned optimal cruising speed. Figure 7 As shown in the figure, the acceleration of the vehicle before and after the control is compared. After the control, the output acceleration is more gentle, which is beneficial to reduce fuel consumption and physiological strain. Figure 8 As shown, the vehicle's fuel consumption dropped from 1.80L to 1.74L, a decrease of 3.45%, improving the vehicle's fuel economy and energy saving.

[0167] In terms of improving physiological strain, such as Figure 9 and Figure 10 The figure below compares the damping forces of the front and rear shock absorbers. The vertical skyhook hybrid control strategy provides a wider damping force adjustment range and faster response. Both vertical and pitch vibrations are alleviated, as shown in Table 3 below. The peak and RMS values ​​for both decrease by an average of 19.15% and 7.68%, respectively.

[0168] Table 3 Vibration statistics

[0169]

[0170] like Figure 11 As shown, human physiological strain indicators It was reduced by 15.38%, which effectively verified the effectiveness of the proposed predictive adaptive cruise control method in improving physiological strain.

[0171] The step numbers in the above embodiments are only provided for the convenience of explanation and do not limit the order of the steps. The execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0172] Example 2

[0173] This embodiment provides a commercial vehicle cruise control system that takes into account both physiological strain and energy conservation, and is used to implement the commercial vehicle cruise control method that takes into account both physiological strain and energy conservation of Example 1, comprising: a human physiological strain index module, a vehicle fuel consumption and energy conservation index module, an optimal cruise speed planning module, a reinforcement learning solution module, a dynamic model module, a longitudinal model predictive control module, and a hybrid control module;

[0174] In this embodiment, the human physiological strain index construction module is used to construct the human physiological strain index;

[0175] In this embodiment, the vehicle fuel consumption and energy saving index construction module is used to construct the vehicle fuel consumption and energy saving index;

[0176] In this embodiment, the optimal cruising speed planning module is used to construct an optimal cruising speed planning model and construct an overall reward function based on the vehicle fuel consumption and energy saving index and travel time;

[0177] In this embodiment, the reinforcement learning solution module is used to optimize the reinforcement learning algorithm and solve the optimal cruising speed. The state, acceleration, and reward value are stored in the predicted experience playback area. During the training process, samples are taken from the predicted experience playback area to update the policy network and value network of the reinforcement learning algorithm within the prediction time domain, and solve the optimal cruising speed within the prediction time.

[0178] In this embodiment, the dynamic model construction module is used to construct a longitudinal and vertical coupled dynamic model of the vehicle-seat-human body system. Acceleration is calculated based on the optimal cruising speed. The acceleration and road excitation are input into the longitudinal and vertical coupled dynamic model of the vehicle-seat-human body system to obtain vertical vibration acceleration and pitch angular acceleration. The human physiological strain index at the optimal cruising speed is then calculated.

[0179] In this embodiment, the longitudinal model predictive control module is used to encode the vehicle state constraints and the optimal cruising speed sequence tracking constraints as control constraints, and form a longitudinal model predictive control strategy based on human physiological strain indicators, vehicle fuel consumption and energy saving indicators, and tracking deviation;

[0180] In this embodiment, the hybrid control module is used to add a pitch control mechanism to realize vertical skyhook hybrid control and regulate the vehicle suspension.

[0181] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be considered as equivalent replacement methods and are included in the scope of protection of the present invention.

Claims

1. A cruise control method for commercial vehicles that takes into account both physiological strain and energy saving, characterized in that: The steps include: Construct human physiological strain indicators, construct vehicle fuel consumption and energy-saving indicators, establish an optimal cruising speed planning model, and construct an overall reward function based on vehicle fuel consumption and energy-saving indicators and travel time; The solution is based on the reinforcement learning algorithm. The state, acceleration, and reward values ​​are stored in the predicted experience playback area. During the training process, samples are taken from the predicted experience playback area to update the policy network and value network of the reinforcement learning algorithm in the prediction time domain, and solve the optimal cruising speed within the prediction time. A longitudinal and vertical coupled dynamic model of the vehicle-seat-human system was constructed. Acceleration was calculated based on the optimal cruising speed. The acceleration and road excitation were input into the longitudinal and vertical coupled dynamic model of the vehicle-seat-human system to obtain vertical vibration acceleration and pitch angular acceleration. The physiological strain index of the human body at the optimal cruising speed was calculated. The vehicle state constraints and optimal cruising speed sequence tracking constraints are encoded as control constraints, and a longitudinal model predictive control strategy is formed based on human physiological strain indicators, vehicle fuel consumption and energy saving indicators, and tracking deviation. The additional pitch control mechanism realizes vertical skylight hybrid control and regulates the vehicle suspension.

2. The commercial vehicle cruise control method that takes into account both physiological strain and energy saving according to claim 1 is characterized in that: Construct human physiological strain indicators, including: The vibration excitation is transformed into frequency domain and weighted. In the frequency domain transformation process, the vertical vibration acceleration is transformed into , pitch vibration acceleration Transformed into vertical vibration acceleration spectrum function , pitch vibration acceleration spectrum function In the weighted processing process, corresponding weighting functions are provided for vibrations in different axes. The weighted vibration acceleration spectrum function is: ； ； in, 、 are the vertical weighted vibration acceleration spectrum function and the pitch weighted vibration acceleration spectrum function, is the vertical vibration weighting function, is the pitch vibration weighting function, 、 is the corresponding weight coefficient; Based on the inverse Fourier transform 、 Transformed into vertical weighted vibration acceleration , pitch weighted vibration acceleration ; The quantitative calculation formula for the impact of vertical and pitch vibration excitation on the human body is constructed as follows: ； ； in, represents the vertical weighted vibration impact value, represents the pitch weighted vibration impact value, Indicates duration; The human body physiological strain index is expressed as: ； in, Indicates the physiological strain index of the human body. 、 、 、 is the model coefficient.

3. The commercial vehicle cruise control method that takes into account both physiological strain and energy saving according to claim 1 is characterized in that: Construct vehicle fuel consumption and energy saving indicators, specifically expressed as: ； ； in, Indicates the vehicle's fuel consumption and energy saving index. Indicates the engine output power, 、 、 are coefficients to be fitted, is the engine speed, is the vehicle speed, is the frontal area of ​​the vehicle, is the drag coefficient, is the vehicle gravity, is the road slope, is the vehicle weight, is the longitudinal acceleration, is the rolling resistance coefficient, The vehicle transmission efficiency.

4. The commercial vehicle cruise control method that takes into account both physiological strain and energy saving according to claim 1 is characterized in that: Establish an optimal cruising speed planning model, including: The iterative process of the predictive cruise speed planning strategy is expressed in discrete form as follows: ； in, For the The vehicle position during step planning, For the The vehicle speed during step planning, For the The vehicle position during step planning, For the The vehicle speed during step planning, For the Vehicle acceleration during step planning, For the The air resistance acceleration caused by air resistance during step planning, is the acceleration due to gravity, is the friction coefficient, For the The road slope corresponding to the vehicle's position during step planning, The planning step duration.

5. The commercial vehicle cruise control method that takes into account both physiological strain and energy saving according to claim 4 is characterized in that: The overall reward function is constructed based on the vehicle fuel consumption and travel time, which is expressed as: ； in, and is the adjustment coefficient, Indicates the vehicle's fuel consumption and energy-saving index.

6. The commercial vehicle cruise control method that takes into account both physiological strain and energy saving according to claim 1 is characterized in that: Solving based on reinforcement learning algorithm, the specific steps include: Add a prediction experience playback area for the TD3 reinforcement learning algorithm And store the state, acceleration, and reward value in the predicted experience playback area, and form an array based on the state, acceleration, and reward value ,in, Indicates the The state of the step solution, is the corresponding acceleration, is the reward value, is the next state; In the policy network and value network of the TD3 reinforcement learning algorithm, the policy network is used to receive current state data, combine and return features to output continuous actions, and the value network obtains sensor input and the action output of the policy network in the current state, and outputs the optimal value of the current state action; In each planning step, the overall reward function is calculated, and the motion parameters for the next moment are calculated by combining the current state and the cruise speed planning model; Based on the spline interpolation algorithm, intermediate points are inserted between adjacent planning steps to obtain the optimal cruising speed within the prediction time.

7. The commercial vehicle cruise control method that takes into account both physiological strain and energy saving according to claim 1 is characterized in that: Construct a vertical and vertical coupled dynamic model of the vehicle-seat-human system, which can be expressed as follows: ； in: 、 、 、 、 、 、 、 、 、 ； in, are the moments of inertia of each part, are the vertical vibration accelerations of each part, are the vertical displacements of each part, and are the vertical displacements of the front and rear suspensions of each part, and are the vertical velocities of the front and rear suspensions of each part, are the pitch displacements of each part, are the pitch angular accelerations of each part, For the quality of each part, is the height of the centroid of each part, and are the front and rear suspension spring stiffness of each part respectively, and are the damping coefficients of the front and rear suspension of each part, and are the distances from the center of mass of each part to the front and rear suspensions, Indicates the pitch angle of each part, subscript Corresponding to the vehicle frame, cab and seat-human body system respectively, and are the displacements of the front and rear suspension lower ends of the seat, and are the displacements of the front and rear suspension lower ends of the cab, and are the distances from the cab center of mass to the front and rear suspensions of the seat, and are the distances from the center of mass of the frame to the front and rear suspensions of the cab, respectively.

8. The commercial vehicle cruise control method that takes into account both physiological strain and energy saving according to claim 1 is characterized in that: The vehicle state constraints are expressed as: ； in, and The minimum and maximum limits of vehicle acceleration are: and The minimum and maximum vehicle speed limits are: is the actual distance between the vehicle and the preceding vehicle, is the minimum safe distance between vehicles. is the time distance between the vehicle and the preceding vehicle; Based on the tracking deviation between the actual vehicle speed and the planned optimal cruising speed Construct a tracking constraint, expressed as: ； in, is the vehicle speed error, is the acceleration error, is the acceleration error weight coefficient; The objective function of the longitudinal model predictive control strategy is expressed as: ； in, 、 、 is the weight factor, Indicates the physiological strain index of the human body. Indicates the vehicle's fuel consumption and energy-saving index.

9. The commercial vehicle cruise control method that takes into account both physiological strain and energy saving according to claim 1, characterized in that: Additional pitch control mechanism realizes vertical ceiling hybrid control, including: The vertical ceiling control strategy is designed in sections according to different working conditions, which can be expressed as: ； in, represents the damping coefficient, represents the vertical acceleration, represents the vertical acceleration threshold, represents the maximum damping coefficient, represents the minimum damping coefficient, Indicates the physiological strain index of the human body; The additional pitch control strategy is: ； in, and are the front and rear seat suspension damping coefficients, is the damping torque distribution coefficient of the front and rear seat suspension, is the distance between the front and rear seat suspensions, is the vertical velocity of the front suspension of the seat-body system, is the vertical velocity of the rear suspension of the seat-body system, Indicates the pitch angular velocity.

10. A cruise control system for commercial vehicles that takes into account both physiological strain and energy saving, characterized in that: A commercial vehicle cruise control method for achieving both physiological strain and energy saving as described in any one of claims 1 to 9, comprising: a human physiological strain index module, a vehicle fuel consumption and energy saving index module, an optimal cruise speed planning module, a reinforcement learning solution module, a dynamic model module, a longitudinal model predictive control module, and a hybrid control module; The human body physiological strain index construction module is used to construct the human body physiological strain index; The vehicle fuel consumption and energy saving index construction module is used to construct the vehicle fuel consumption and energy saving index; The optimal cruising speed planning module is used to construct an optimal cruising speed planning model and build an overall reward function based on vehicle fuel consumption and energy saving indicators and travel time; The reinforcement learning solution module is used to optimize the reinforcement learning algorithm and solve the optimal cruising speed. The state, acceleration, and reward value are stored in the predicted experience playback area. During the training process, samples are taken from the predicted experience playback area to update the policy network and value network of the reinforcement learning algorithm in the prediction time domain, and solve the optimal cruising speed within the prediction time. The dynamic model building module is used to build a longitudinal and vertical coupled dynamic model of the vehicle-seat-human body system, calculate acceleration based on the optimal cruising speed, input the acceleration and road excitation into the longitudinal and vertical coupled dynamic model of the vehicle-seat-human body system to obtain vertical vibration acceleration and pitch angular acceleration, and calculate the human physiological strain index at the optimal cruising speed; The longitudinal model predictive control module is used to encode the vehicle state constraints and the optimal cruising speed sequence tracking constraints as control constraints, and form a longitudinal model predictive control strategy based on human physiological strain indicators, vehicle fuel consumption and energy saving indicators, and tracking deviation; The hybrid control module is used to add a pitch control mechanism to realize vertical skyhook hybrid control and regulate the vehicle suspension.

Images