Calibration system and method for accelerator and brake of unmanned tank car with load

Abstract

The invention relates to an accelerator and brake calibration system and method for an unmanned tank car with a load. The method comprises the following steps that running speed data and running acceleration data of the tank car under different load conditions, different accelerator opening degree conditions and different brake opening degree conditions are collected; performing deep neural network model training by using the collected data to obtain an accelerator and brake calibration model; inputting the formulated load parameters, speed parameters, accelerator opening parameters and brake opening parameters into an accelerator and brake calibration model, and further obtaining acceleration parameters output by the accelerator and brake calibration model; and the load parameters, the speed parameters, the accelerator opening parameters, the brake opening parameters and the corresponding acceleration parameters are associated in a one-to-one correspondence mode to form an accelerator and brake calibration table. According to the calibration table, the important dimension of the load is added, so that the calibration is more targeted, and the tank car is more stable and reliable in performance under different load states.

Description

Technical Field

[0001] The present invention relates to the technical field of throttle and brake calibration in unmanned driving, and in particular to a throttle and brake calibration system and method for an unmanned tank truck with a load. Background Art

[0002] Faced with the shortage of cement tanker drivers and the harsh working environment, the research on unmanned cement tankers has become a focus. The most important technology for unmanned cement tankers is the calibration of the throttle and brake. In the existing technology, the tanker chassis is modified to add wire control communication control, and the speed and acceleration data of different throttle and brake positions in the wire control chassis of the cement tanker under different load conditions are collected. The acceleration data is interpolated and corrected, and the throttle and brake calibration table is calculated and generated based on this.

[0003] The existing methods have the following problems:

[0004] (1) Heavy measurement workload: When using the interpolation method to calibrate the throttle and brake, it is necessary to measure multiple throttle positions to ensure the accuracy of the interpolation, and repeat the measurement to reduce the statistical error, resulting in a huge workload on site.

[0005] (2) Load fluctuations are not taken into account: the existing technology is only applicable to fixed loads and does not have the robustness to be adjusted flexibly. However, the load of the cement tanker will change continuously over a large range during use.

[0006] (3) Applicable only to forward gear: The existing technology is only applicable to acceleration measurement and throttle and brake calibration when the vehicle is moving forward, and the same calibration table is applied to reverse operation.

[0007] In summary, the calibration method suitable for passenger cars is not suitable for heavy-duty engineering vehicles. In addition, cement tank trucks and passenger cars have different gears. Using the forward gear calibration table for reverse operation according to the traditional calibration method of passenger cars will significantly affect the vehicle's movement accuracy. Summary of the invention

[0008] The purpose of the present invention is to overcome the defects of the prior art and provide a throttle and brake calibration system and method for an unmanned tank truck with load, so as to solve the problems that the existing passenger car calibration method is not suitable for heavy engineering vehicles, the existing passenger car calibration method has heavy measurement work, does not consider the responsible shifting situation, and is only applicable to the forward gear.

[0009] The technical solution to achieve the above purpose is:

[0010] The present invention provides a method for calibrating the throttle and brake of an unmanned tank truck with a load, comprising the following steps:

[0011] Collect the driving speed data and driving acceleration data of the tank truck under different load conditions, different throttle opening conditions and different brake opening conditions;

[0012] The collected driving speed data, driving acceleration data, corresponding load, throttle opening and brake opening of the tank truck are used to train a deep neural network model to obtain a throttle brake calibration model;

[0013] Inputting the formulated load parameter, speed parameter, throttle opening parameter and brake opening parameter into the throttle brake calibration model, thereby obtaining the acceleration parameter output by the throttle brake calibration model;

[0014] The load parameters, speed parameters, throttle opening parameters, brake opening parameters and corresponding acceleration parameters are associated one by one to form a throttle brake calibration table, thereby completing the calibration of the tank truck throttle and brake.

[0015] The further improvement of the throttle and brake calibration method of the unmanned tank truck with load of the present invention is that it also includes:

[0016] Provide RTK sensors and IMU sensors, and install the provided RTK sensors and IMU sensors on the tank truck;

[0017] Acquiring the driving speed data and driving acceleration data of the tank truck during the travel process through the RTK sensor and the IMU sensor;

[0018] Performing speed differential calculation on the driving speed data according to a set time interval to obtain an acceleration calculation value at a corresponding time point;

[0019] The acceleration calculation value and the corresponding driving acceleration value in the driving acceleration data are fused using an extended Kalman filter algorithm to obtain fused acceleration data.

[0020] The further improvement of the throttle and brake calibration method of the unmanned tank truck with load of the present invention is that the speed difference calculation is performed using the following formula:

[0021]

[0022] In the formula, a t Indicates the acceleration at the target time point, v t and v t-1 are the speeds at the current time point and the previous time point respectively, and Δt is the time interval.

[0023] A further improvement of the throttle and brake calibration method of a loaded unmanned tank truck of the present invention is that when collecting the driving speed data and driving acceleration data of the tank truck under different load conditions, the load of the tank truck is divided into intervals, and independent driving speed data and driving acceleration data are collected for each load interval.

[0024] The further improvement of the throttle and brake calibration method of the unmanned tank truck with load of the present invention is that it also includes:

[0025] A load sensor is provided, and the load sensor is installed on the tank truck. The load sensor is used to detect and record the load condition of the tank truck.

[0026] The present invention also provides a throttle and brake calibration system for an unmanned tank truck with a load, comprising:

[0027] A collection unit, used to collect the driving speed data and driving acceleration data of the tank truck under different load conditions, different throttle opening conditions and different brake opening conditions;

[0028] A model training unit connected to the collection unit is used to perform deep neural network model training using the driving speed data, driving acceleration data, corresponding load, throttle opening and brake opening of the tank truck collected by the collection unit to obtain a throttle brake calibration model;

[0029] An input unit connected to the throttle-brake calibration model, used for inputting the formulated load parameter, speed parameter, throttle opening parameter and brake opening parameter into the throttle-brake calibration model;

[0030] The processing unit connected to the input unit and the throttle-brake calibration model is used to receive the acceleration parameters output by the throttle-brake calibration model, and is also used to associate the load parameters, speed parameters, throttle opening parameters, and brake opening parameters with corresponding acceleration parameters one-to-one to form a throttle-brake calibration table.

[0031] A further improvement of the throttle and brake calibration system of the unmanned tank truck with load of the present invention is that it also includes an RTK sensor and an IMU sensor installed on the tank truck;

[0032] The RTK sensor and the IMU sensor are used to obtain the driving speed data and driving acceleration data during the tank truck travel process;

[0033] The collecting unit is connected to the RTK sensor and the IMU sensor, and collects the driving speed data and driving acceleration data of the tank truck under corresponding conditions through the RTK sensor and the IMU sensor;

[0034] The model training unit is also used to perform speed differential calculation on the driving speed data according to a set time interval to obtain an acceleration calculation value at a corresponding time point, and use an extended Kalman filter algorithm to fuse the acceleration calculation value and the corresponding driving acceleration value in the driving acceleration data to obtain fused acceleration data, and then use the fused acceleration data to perform deep neural network model training.

[0035] A further improvement of the throttle and brake calibration system of the unmanned tank truck with load of the present invention is that a speed difference calculation formula is preset in the model training unit:

[0036]

[0037] In the formula, a t Indicates the acceleration at the target time point, v t and v t-1 are the speeds at the current time point and the previous time point respectively, and Δt is the time interval;

[0038] The speed difference calculation formula is used to perform speed difference calculation.

[0039] A further improvement of the throttle and brake calibration system of the loaded unmanned tank truck of the present invention is that the collection unit is also used to divide the load of the tank truck into intervals, and to collect independent driving speed data and driving acceleration data for each load interval.

[0040] A further improvement of the throttle and brake calibration system for the loaded unmanned tank truck of the present invention is that it also includes a load sensor installed on the tank truck for detecting and recording the load condition of the tank truck.

[0041] The beneficial effects of the throttle and brake calibration system and method of the unmanned tank truck with load are as follows:

[0042] The calibration table of the present invention adds the important dimension of load, making the calibration more targeted and making the performance of the tank truck more stable and reliable under different load conditions.

[0043] The present invention performs fusion processing on acceleration data, and can provide high-precision and low-latency acceleration information, thereby improving the robustness and accuracy of the data.

[0044] The calibration system and method of the present invention utilize a deep neural network model to calibrate the throttle and brake of a tank truck, effectively alleviating the impact of switching between different load states on the throttle and brake calibration.

[0045] The calibration table provided by the calibration system and method of the present invention can automatically select the most appropriate calibration strategy based on the current actual load of the tank truck, thereby greatly improving the adaptability and performance of the vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 The present invention is a flow chart of the throttle and brake calibration method of an unmanned tank truck with load.

[0047] Figure 2 This is the hidden layer and nonlinear mapping relationship diagram of the throttle and brake calibration system and method for the unmanned tank truck with load Zhang Gong trained in the present invention.

[0048] Figure 3 This is a visualization diagram of the throttle and brake calibration system of the unmanned tank truck with load and the acceleration changes inferred by the deep neural network in the system under no-load, 0.5-ton and 1-ton load conditions. DETAILED DESCRIPTION

[0049] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0050] See also Figure 1 The present invention provides a throttle and brake calibration system and method for an unmanned tank truck with load, aiming to improve the speed of vehicle throttle and brake calibration and the accuracy of vehicle control when the load of heavy engineering vehicles changes significantly, so as to meet the power matching requirements. Considering that heavy engineering vehicles such as cement tank trucks have a great impact on vehicle speed and acceleration under the influence of different loads, and it is impossible to enumerate the load conditions of the vehicle in actual calibration, the present invention proposes a method for making a throttle and brake calibration table based on a neural network, which can effectively alleviate the impact on the throttle and brake calibration table under the switching of different states such as load and road surface. The throttle and brake calibration system and method for an unmanned tank truck with load of the present invention are described below in conjunction with the accompanying drawings.

[0051] The present invention provides a throttle and brake calibration system for an unmanned tank truck with a load, which includes a collection unit, a model training unit, an input unit and a processing unit; wherein the model training unit is connected to the collection unit, the input unit is connected to the throttle and brake calibration model, the processing unit is connected to the input unit and the throttle and brake unit, and the collection unit is used to collect the driving speed data and driving acceleration data of the tank truck under different load conditions, different throttle opening conditions and different brake opening conditions. Specifically, the tank truck is allowed to perform corresponding driving tests according to the load conditions, throttle conditions and brake conditions, and the corresponding data is collected during the driving test, including the driving conditions of the tank truck (the above-mentioned load conditions, throttle conditions and brake conditions), driving speed and driving acceleration. Further, when the tank truck is allowed to perform a driving test, different gear conditions and different road conditions are added to achieve multi-dimensional calibration of the driving speed of the tank truck.

[0052] The model training unit is used to perform deep neural network model training using the tank truck's driving speed data, driving acceleration data, corresponding load, throttle opening, and brake opening collected by the collection unit to obtain a throttle brake calibration model. When performing deep neural network model training, the tank truck's driving speed data, corresponding load, throttle opening, and brake opening conditions are used as model inputs, and the driving acceleration data is used as output, such as Figure 2 As shown, a deep neural network has multiple hidden layers to achieve complex nonlinear mapping. Each hidden layer consists of multiple neurons, and each neuron has a weight and a bias parameter.

[0053] During training, the model adjusts these parameters through two steps: forward propagation and backpropagation.

[0054] In the forward propagation phase, the model makes predictions based on the input data and the current weight parameters. The output of each neuron is the product of its input and weight plus the bias, which is then transformed nonlinearly through the activation function.

[0055] In the back-propagation phase, the model updates the weights and biases based on the difference between the predicted results and the true values, i.e. the loss. By calculating the gradient of the loss function with respect to each parameter, the model can know how to adjust the parameters to reduce the prediction error.

[0056] This process is repeated for multiple training cycles until the model's prediction error on the training data reaches an acceptable level, or until a preset number of training cycles is reached. In this way, the model is able to learn the complex mapping relationship from input data to prediction results.

[0057] The input unit is used to input the formulated load parameter, speed parameter, throttle opening parameter and brake opening parameter into the throttle brake calibration model;

[0058] The processing unit is used to receive the acceleration parameters output by the throttle brake calibration model, and is also used to associate the load parameters, speed parameters, throttle opening parameters, brake opening parameters with the corresponding acceleration parameters one by one to form a throttle brake calibration table. Further, the processing unit associates and stores the load parameters, speed parameters, throttle opening parameters, brake opening parameters with the corresponding acceleration parameters in the form of a table.

[0059] The throttle and brake calibration table of the present invention has four dimensions, including load, speed, acceleration and instructions (throttle instructions and brake instructions). The process of using the throttle and brake calibration table in the automatic unmanned driving of the tank truck is to obtain the corresponding instructions according to the current speed, load and expected acceleration. The instructions can be throttle instructions or brake instructions. The obtained instructions are sent to the control system of the vehicle so that the corresponding components can execute them, thereby realizing automatic control of the vehicle.

[0060] In a specific embodiment of the present invention, the system of the present invention further comprises an RTK sensor and an IMU sensor installed on the tank truck; the RTK sensor and the IMU sensor are used to obtain the driving speed data and the driving acceleration data during the tank truck travel process;

[0061] The collection unit of the system of the present invention is connected to the RTK sensor and the IMU sensor, and collects the driving speed data and driving acceleration data of the tank truck under corresponding conditions through the RTK sensor and the IMU sensor;

[0062] The model training unit is also used to perform speed differential calculation on the driving speed data according to a set time interval to obtain the acceleration calculation value at the corresponding time point, and use the extended Kalman filter algorithm to fuse the acceleration calculation value and the corresponding driving acceleration value in the driving acceleration data to obtain the fused acceleration data, and then use the fused acceleration data to perform deep neural network model training.

[0063] Furthermore, a speed difference calculation formula is preset in the model training unit:

[0064]

[0065] In the formula, a t Indicates the acceleration at the target time point, v t and v t-1 are the speeds at the current time point and the previous time point respectively, and Δt is the time interval;

[0066] The speed difference calculation formula is used to perform speed difference calculation.

[0067] The present invention integrates the acceleration data (a t) and the data measured by the IMU sensor (a′ t ), the system of the present invention uses an extended Kalman filter (EKF) to fine-tune the vehicle speed estimate. In this process, the core role of the EKF is reflected in two aspects: first, it integrates two data sources to increase the amount of information in the estimate; second, through its ability to handle nonlinearity, it effectively reduces noise and errors in the measurement, thereby improving the accuracy and stability of the speed estimate, significantly reducing the uncertainty caused by measurement noise and random disturbances, and further enhancing the accuracy and reliability of the system's speed estimate.

[0068] Furthermore, it also includes a load sensor installed on the tank truck, which is used to detect and record the load condition of the tank truck.

[0069] The collection unit of the present invention is also connected to the load sensor, and is used to receive the load data of the tank truck detected by the load sensor, and to form collected data according to the load data, speed data, acceleration data and corresponding throttle opening and brake opening at the corresponding moment, and send it to the model training unit for model training.

[0070] Furthermore, when the vehicle is automatically driving, the load sensor can also monitor and accurately record the current load condition of the vehicle in real time, and use these data as input signals. The system of the present invention can select the best throttle and brake configuration that matches the current load from the calibration table based on these data as input signals, ensuring that the vehicle can maintain optimal performance under various operating conditions.

[0071] In a specific embodiment of the present invention, the collection unit is also used to divide the load of the tank truck into intervals, and collect independent driving speed data and driving acceleration data for each load interval. When dividing the load, the load is divided into intervals from no load to full load to form multiple load intervals, such as 0-0.5 tons, 0.5-1 tons, 1-1.5 tons, etc., and independent data collection and precise calibration are performed for the Mengo interval, or several specific load values ​​can be selected in each load interval for corresponding vehicle driving tests to improve the accuracy of vehicle protection.

[0072] The present invention inputs specific load, speed and command (accelerator or brake) into the deep neural network model to obtain the corresponding acceleration output. Therefore, we create a four-dimensional table containing load, speed, acceleration and command, and use the deep neural network to infer and predict the speed and acceleration changes of the unmanned cement tanker under different loads during driving. The corresponding visualization results under no load, 0.5 ton and 1 ton load are as follows: Figure 3 shown.

[0073] The engine of the cement tank truck of the present invention is electrically driven, but the linearity of the opening percentage of the motor and the movement speed of the mechanism is not good, and it is necessary to compensate for this nonlinear effect by establishing a lookup table.

[0074] The present invention also provides a method for calibrating the throttle and brake of an unmanned tank truck with load, and the calibration method is described below.

[0075] like Figure 1 As shown, the calibration method of the present invention comprises the following steps:

[0076] S11, data collection and processing, collecting the driving speed data and driving acceleration data of the tank truck under different load conditions, different throttle opening conditions and different brake opening conditions;

[0077] S12, training a deep neural network model, using the collected driving speed data, driving acceleration data, corresponding load, throttle opening, and brake opening of the tank truck to train the deep neural network model to obtain a throttle brake calibration model;

[0078] S13, preparing a throttle-brake calibration table, inputting the formulated load parameter, speed parameter, throttle opening parameter and brake opening parameter into the throttle-brake calibration model, and then obtaining the acceleration parameter output by the throttle-brake calibration model;

[0079] The load parameters, speed parameters, throttle opening parameters, brake opening parameters and corresponding acceleration parameters are associated one by one to form a throttle and brake calibration table, thereby completing the calibration of the tank truck throttle and brake.

[0080] In a specific embodiment of the present invention, the calibration method of the present invention further includes:

[0081] Provide RTK sensors and IMU sensors, and install the provided RTK sensors and IMU sensors on the tank truck;

[0082] The driving speed data and driving acceleration data of the tank truck during the journey are obtained through the RTK sensor and IMU sensor;

[0083] Perform speed differential calculation on the driving speed data according to a set time interval to obtain the acceleration calculation value at the corresponding time point;

[0084] The extended Kalman filter algorithm is used to fuse the acceleration calculation value and the corresponding driving acceleration value in the driving acceleration data to obtain fused acceleration data.

[0085] Furthermore, the velocity difference is calculated using the following formula:

[0086]

[0087] In the formula, a t Indicates the acceleration at the target time point, v t and v t -1 are the speeds at the current time point and the previous time point respectively, and Δt is the time interval.

[0088] The data collection of the present invention needs to collect a large amount of vehicle driving data under different loads, different throttle positions and braking forces. This includes driving performance in different gears (including reversing) and different road conditions. It is specifically responsible for collecting and processing data of the vehicle during movement: at each measurement interval, the Real-Time kinematics (RTK) sensor and the Inertial-measurement unit (IMU) sensor will automatically record the speed at that time point, and combine the data of the previous time interval to calculate the acceleration of the vehicle at the current gear and speed. After further alignment, the data from the two sources are fused through the extended Kalman filter algorithm. This fusion method can provide high-precision and low-latency acceleration information, thereby improving the robustness and accuracy of the data. The load data should include all stages from no load to full load to ensure the diversity and comprehensiveness of the data.

[0089] Collect data: The goal of this phase is to reduce the data noise and errors from RTK and IMU sources to provide clean data input for data fusion.

[0090] Processing data:

[0091] Speed ​​difference calculation: Select an appropriate time interval (Δt) for the speed information and calculate the acceleration (a t ), the calculation method is (v t ) and (v t-1 ) are the speeds at the current and previous time points respectively.

[0092] Acceleration reading: Read the acceleration information in the sensor and determine the acceleration (a′) at the target time point t ) is used to eliminate a certain delay in velocity difference calculation in data fusion.

[0093] Acceleration data fusion: The acceleration data calculated by the fusion velocity sensor (a t ) and the data directly measured by the accelerometer (a′ t), this system uses the extended Kalman filter (EKF) to fine-tune the vehicle speed estimation. In this process, the core role of EKF is reflected in two aspects: first, it integrates two data sources to increase the amount of information in the estimation; second, through its nonlinear processing ability, it effectively reduces the noise and error in the measurement, thereby improving the accuracy and stability of speed estimation, significantly reducing the uncertainty caused by measurement noise and random disturbances, and further enhancing the accuracy and reliability of the system's speed estimation.

[0094] In a specific embodiment of the present invention, when collecting the driving speed data and driving acceleration data of the tank truck under different load conditions, the load of the tank truck is divided into intervals, and independent driving speed data and driving acceleration data are collected for each load interval.

[0095] In a specific embodiment of the present invention, it also includes:

[0096] A load sensor is provided, and the load sensor is installed on the tank truck. The load sensor is used to detect and record the load condition of the tank truck.

[0097] The present invention is described in detail above in conjunction with the embodiments of the accompanying drawings. A person skilled in the art can make various variations of the present invention according to the above description. Therefore, certain details in the embodiments should not constitute a limitation of the present invention, and the scope of protection of the present invention shall be defined by the scope of the attached claims.

Claims

1. A method for calibrating the throttle and brake of an unmanned tank truck with load, characterized in that: The steps include: Collect the driving speed data and driving acceleration data of the tank truck under different load conditions, different throttle opening conditions and different brake opening conditions; The collected driving speed data, driving acceleration data, corresponding load, throttle opening and brake opening of the tank truck are used to train a deep neural network model to obtain a throttle brake calibration model; Inputting the formulated load parameter, speed parameter, throttle opening parameter and brake opening parameter into the throttle brake calibration model, thereby obtaining the acceleration parameter output by the throttle brake calibration model; The load parameters, speed parameters, throttle opening parameters, brake opening parameters and corresponding acceleration parameters are associated one by one to form a throttle brake calibration table, thereby completing the calibration of the tank truck throttle and brake.

2. The throttle and brake calibration method for an unmanned tank truck with load as claimed in claim 1 is characterized in that: Also includes: Provide RTK sensors and IMU sensors, and install the provided RTK sensors and IMU sensors on the tank truck; Acquiring the driving speed data and driving acceleration data of the tank truck during the travel process through the RTK sensor and the IMU sensor; Performing speed differential calculation on the driving speed data according to a set time interval to obtain an acceleration calculation value at a corresponding time point; The acceleration calculation value and the corresponding driving acceleration value in the driving acceleration data are fused using an extended Kalman filter algorithm to obtain fused acceleration data.

3. The throttle and brake calibration method for an unmanned tank truck with load as claimed in claim 2 is characterized in that: The speed difference is calculated using the following formula: In the formula, a t Indicates the acceleration at the target time point, v t and v t-1 are the speeds at the current time point and the previous time point respectively, and Δt is the time interval.

4. The throttle and brake calibration method for an unmanned tank truck with load as claimed in claim 1 is characterized in that: When collecting the driving speed data and driving acceleration data of the tank truck under different load conditions, the load of the tank truck is divided into intervals, and independent driving speed data and driving acceleration data are collected for each load interval.

5. The throttle and brake calibration method for an unmanned tank truck with load as claimed in claim 1 is characterized in that: Also includes: A load sensor is provided, and the load sensor is installed on the tank truck. The load sensor is used to detect and record the load condition of the tank truck.

6. A throttle and brake calibration system for an unmanned tank truck with load, characterized in that: include: A collection unit, used to collect the driving speed data and driving acceleration data of the tank truck under different load conditions, different throttle opening conditions and different brake opening conditions; A model training unit connected to the collection unit is used to perform deep neural network model training using the driving speed data, driving acceleration data, corresponding load, throttle opening and brake opening of the tank truck collected by the collection unit to obtain a throttle brake calibration model; An input unit connected to the throttle-brake calibration model, used for inputting the formulated load parameter, speed parameter, throttle opening parameter and brake opening parameter into the throttle-brake calibration model; The processing unit connected to the input unit and the throttle-brake calibration model is used to receive the acceleration parameters output by the throttle-brake calibration model, and is also used to associate the load parameters, speed parameters, throttle opening parameters, and brake opening parameters with corresponding acceleration parameters one-to-one to form a throttle-brake calibration table.

7. The throttle and brake calibration system for an unmanned tank truck with load as claimed in claim 6 is characterized in that: It also includes RTK sensors and IMU sensors installed on tank trucks; The RTK sensor and the IMU sensor are used to obtain the driving speed data and driving acceleration data during the tank truck travel process; The collecting unit is connected to the RTK sensor and the IMU sensor, and collects the driving speed data and driving acceleration data of the tank truck under corresponding conditions through the RTK sensor and the IMU sensor; The model training unit is also used to perform speed differential calculation on the driving speed data according to a set time interval to obtain an acceleration calculation value at a corresponding time point, and use an extended Kalman filter algorithm to fuse the acceleration calculation value and the corresponding driving acceleration value in the driving acceleration data to obtain fused acceleration data, and then use the fused acceleration data to perform deep neural network model training.

8. The throttle and brake calibration system for an unmanned tank truck with load as claimed in claim 6, characterized in that: The model training unit is preset with a speed difference calculation formula: In the formula, a t Indicates the acceleration at the target time point, v t and v t-1 are the speeds at the current time point and the previous time point respectively, and Δt is the time interval; The speed difference calculation formula is used to perform speed difference calculation.

9. The throttle and brake calibration system for an unmanned tank truck with load as claimed in claim 6, characterized in that: The collection unit is also used to divide the load of the tank truck into intervals, and collect independent driving speed data and driving acceleration data for each load interval.

10. The throttle and brake calibration system for an unmanned tank truck with load as claimed in claim 6, characterized in that: It also includes a load sensor installed on the tank truck, which is used to detect and record the load condition of the tank truck.

Images