A braking detection method for an electric tricycle

Abstract

The present invention discloses a braking detection method for an electric tricycle, which relates to the technical field of braking detection. The method includes: collecting a historical braking accident data set with the type of the electric tricycle to be detected as a constraint; obtaining the load constraint characteristics of the electric tricycle to be detected; performing effective auxiliary data screening to generate an effective braking accident data set; configuring the road environment and load based on the effective braking accident data set to establish an effective test scenario; and performing a combined test on the braking timeliness and braking stability of the electric tricycle to be detected based on the effective test scenario to generate a braking detection result of the electric tricycle to be detected. The present invention solves the technical problems in the prior art that there are large deviations in the evaluation results of the braking performance of electric tricycles and it is difficult to accurately reflect the actual working conditions, and achieves the technical effect of improving the adaptability and result accuracy of the braking performance evaluation of electric tricycles to the actual operating conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of braking detection, and particularly relates to a braking detection method for an electric tricycle. Background Art

[0002] In the field of research on the braking performance of electric tricycles, traditional testing methods often lack careful consideration of the diversity of vehicle types and actual load conditions. Generally, the collection of test data is rather general. Instead of collecting specific historical braking accident data for a particular type of electric tricycle, broad and general data sources are used, resulting in a low degree of fit between the data and the actual situation of the vehicle to be detected. At the same time, in terms of load, only the general load-bearing range of the vehicle is simply considered, without in-depth analysis of the weight of passengers in the front seat and the specific constraint characteristics of the cargo in the cargo compartment, such as the influence of different numbers and weight distributions of passengers in the front seat and different types of goods and changes in the center of gravity position in the cargo compartment on the braking performance. This rough data collection and analysis method makes it impossible to accurately simulate various working conditions of the vehicle during actual operation when conducting braking tests, resulting in large errors and uncertainties in the braking performance evaluation results, making it difficult to effectively guide the optimization and improvement of the vehicle braking system and unable to meet the growing safety requirements for the driving of electric tricycles.

[0003] The prior art has the technical problems of a large deviation in the evaluation result of the braking performance of electric tricycles and difficulty in accurately reflecting the actual working conditions. Summary of the Invention

[0004] The present application provides a braking detection method for an electric tricycle, which is used to solve the technical problems in the prior art of a large deviation in the evaluation result of the braking performance of electric tricycles and difficulty in accurately reflecting the actual working conditions.

[0005] In view of the above problems, the present application provides a braking detection method for an electric tricycle, and the method includes:

[0006] Taking the type of the electric tricycle to be detected as a constraint, collecting a historical braking accident data set; obtaining the load constraint characteristics of the electric tricycle to be detected, where the load constraint characteristics include a front-seat load constraint and a cargo-compartment load constraint; based on the front-seat load constraint and the cargo-compartment load constraint, screening effective auxiliary data from the historical braking accident data set to generate an effective braking accident data set; configuring the road environment and load based on the effective braking accident data set to establish an effective test scenario; and based on the effective test scenario, jointly testing the braking timeliness and braking stability of the electric tricycle to be detected to generate the braking detection result of the electric tricycle to be detected.

[0007] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0008] Taking the type of the electric tricycle to be detected as a constraint, collect a historical braking accident dataset; obtain the load constraint characteristics of the electric tricycle to be detected; screen the effective auxiliary data from the historical braking accident dataset to generate an effective braking accident dataset; configure the road environment and load to establish an effective test scenario; based on the effective test scenario, conduct a joint test on the braking timeliness and braking stability of the electric tricycle to be detected to generate the braking detection result of the electric tricycle to be detected. The technical effect of improving the adaptability of the braking performance evaluation of the electric tricycle to the actual operating conditions and the accuracy of the result is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0010] Figure 1 It is a schematic flowchart of a braking detection method for an electric tricycle provided by an embodiment of the present application.

[0011] Figure 2 It is a schematic flowchart of establishing an effective test scenario in a braking detection method for an electric tricycle provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0012] The present application provides a braking detection method for an electric tricycle to solve the technical problems in the prior art that there are large deviations in the evaluation results of the braking performance of electric tricycles and it is difficult to accurately reflect the actual working conditions.

[0013] The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts belong to the scope of protection of the present application.

[0014] Embodiment, as Figure 1 shown, the present application provides a braking detection method for an electric tricycle, and the method includes:

[0015] Step S100: Taking the type of the electric tricycle to be detected as a constraint, collect a historical braking accident dataset.

[0016] Specifically, clarify the specific type of the electric tricycle to be detected, including key information such as the vehicle model, size, designed use, and the specifications of the braking system equipped. These information will serve as important constraints for collecting the historical braking accident dataset. Then, collect data through various channels. For example, cooperate with the traffic management department to obtain the detailed information of various braking accidents of this type of electric tricycle recorded by them in the past; at the same time, widely consult the claim settlement data of relevant insurance companies and screen out the case information of braking accidents involving this type of vehicle; in addition, it is also possible to conduct in-depth research on the repair records of vehicle repair factories and collect the accident details of this type of vehicle that enters the factory for repair due to braking problems. During the collection process, not only record the basic information such as the time, location, and weather conditions of the accident, but also record in detail the key data such as the driving speed, load condition, braking operation method of the vehicle at the time of the accident, and the consequences caused by the accident, so as to construct a comprehensive, rich and closely related historical braking accident dataset for the type of electric tricycle to be detected, providing a solid data basis for subsequent detection and analysis.

[0017] Step S200: Obtain the load constraint characteristics of the electric tricycle to be detected, where the load constraint characteristics include the front seat load constraint and the cargo box load constraint.

[0018] Specifically, for the electric tricycle to be detected, obtaining its load constraint characteristics is a key link. For the front seat load constraint, it is necessary to carefully consider the number of passengers that can be carried and the per capita weight standard specified in the vehicle design, and at the same time combine different situations such as full load and partial load that may occur in the actual use scenario to determine the influence range of the front seat on the braking performance of the vehicle under different load conditions. This includes analyzing factors such as the distribution method of front seat passengers, sitting posture changes, and possible center of gravity offsets. For the cargo box load constraint, it is necessary to accurately measure the volume of the cargo box and the upper limit of the carrying weight, and investigate the types of goods, weight distribution patterns, and cargo loading and unloading methods of this type of electric tricycle in actual applications such as common logistics distribution and cargo transportation. For example, whether it is evenly distributed loading or concentrated loading in a certain area. These different load conditions will have a significant impact on the center of gravity, inertia, etc. of the vehicle, and thus affect the braking effect. Through a comprehensive and detailed analysis of the front seat load constraint and the cargo box load constraint, the changing trend of the braking performance of the vehicle under various load conditions can be accurately grasped, providing a key basis for subsequent data screening and test scenario construction.

[0019] Step S300: Based on the front seat load constraint and the cargo box load constraint, perform effective auxiliary data screening on the historical braking accident dataset to generate an effective braking accident dataset.

[0020] Specifically, based on the determined front seat load constraint and cargo compartment load constraint, precise and effective auxiliary data screening work is carried out on the collected historical braking accident dataset. First, check one by one the front seat passenger situation and cargo details in the cargo compartment of the vehicle recorded in each piece of data in the historical braking accident dataset, and carefully compare them with the number of people and weight range of the front seat load constraint, and the load capacity and distribution characteristics of the cargo compartment load constraint. For example, if the front seat load constraint is specified as a maximum of two adults, with an average weight of 70 kilograms per person, and the cargo compartment load constraint is a maximum load of 500 kilograms with uniform distribution loading, then filter out the actual accident data that meets this load condition or is similar to it. At the same time, considering factors such as the vehicle's driving speed, road slope, and braking operation details under these load conditions, further exclude the accident data caused by special situations that are significantly different from the conventional usage scenarios. Through this rigorous screening process, the data that does not meet the load constraint conditions is eliminated, thus generating an effective braking accident dataset that can truly reflect the braking performance of the electric tricycle to be detected under normal load conditions, providing highly targeted data support for establishing an effective test scenario in the follow-up.

[0021] Step S400: Based on the effective braking accident dataset, configure the road environment and load to establish an effective test scenario.

[0022] Specifically, extract M pieces of braking accident data from the effective braking accident dataset, each containing the accident road environment characteristics and braking accident characteristics. Then, perform clustering and scenario reference value analysis on these M pieces of data. By identifying the data with a road environment similarity greater than a preset value for clustering to obtain N clustering clusters, calculate the proportion coefficient of the data within each cluster in the M pieces of data, analyze the distribution state of the braking accident characteristics to establish an accident typical coefficient, and then weight the proportion coefficient and the accident typical coefficient to obtain the scenario value degree. Then, filter out the effective accident road environment characteristics corresponding to the clustering clusters with a scenario value degree greater than the preset threshold. Finally, based on these effective accident road environment characteristics and considering the front seat load constraint and the cargo compartment load constraint, precisely configure the road environment (such as slope, road surface condition, etc.) and vehicle load (number of people and weight in the front seat, cargo situation in the cargo compartment) in the test scenario, thus establishing an effective test scenario.

[0023] Step S500: Based on the effective test scenario, conduct a joint test on the braking timeliness and braking stability of the electric tricycle to be detected, and generate the braking detection result of the electric tricycle to be detected.

[0024] Specifically, place the electric tricycle to be tested in an effective test scenario, and start the joint test of its braking timeliness and braking stability to obtain the braking test result. First, let the vehicle brake after reaching the predetermined test speed, and at the same time use the sensing module to collect braking response data to form a braking test data set. Then, extract the braking distance and braking response time from this data set to obtain the braking timeliness test result. Next, analyze the braking force distribution data of the front, rear, left, and right wheels in the data set, compare it with the balanced braking force distribution characteristics to determine the braking force offset vector, and obtain the braking force balance test result. It is also necessary to extract the vehicle body motion attitude data, analyze its pitch, roll, and lateral motion postures, and obtain the braking stability test result. Finally, comprehensively combine the braking timeliness, braking force balance, and braking stability test results to generate the braking test result of the electric tricycle to be tested.

[0025] In a possible implementation manner, as Figure 2 shown, step S400 further includes:

[0026] Step S410: Extract M pieces of braking accident data from the effective braking accident data set, where M is the total number of accidents corresponding to the effective braking accident data set. Any one piece of braking accident data includes accident road environment characteristics and braking accident characteristics.

[0027] Step S420: Based on the accident road environment characteristics and braking accident characteristics, analyze the clustering and scene reference value of the M pieces of braking accident data to obtain N clustering clusters and N scene value degrees, where N is a positive integer less than M.

[0028] Step S430: Based on the N clustering clusters and the N scene value degrees, screen the effective accident road environment characteristics corresponding to the clustering clusters with scene value degrees greater than the preset value threshold.

[0029] Step S440: Configure the test scenario with the effective accident road environment characteristics to establish the effective test scenario.

[0030] Specifically, deeply mine the effective braking accident data set. Determine the value of M as the total number of accidents covered by this data set, and then extract M pieces of braking accident data from it. Each piece of braking accident data contains important information. Among them, the accident road environment characteristics cover the slope conditions of the road, such as steep slopes, gentle slopes, or flat ground; the road surface smoothness characteristics, such as dry, wet, smooth (possibly due to water accumulation, oil stains, etc.); and weather characteristics, such as sunny, rainy, snowy, foggy days, etc. The braking accident characteristics reflect the severity of the problems that occur during braking, including the increase in braking distance, which may exceed the normal range by how much; the angle and direction of the side bending of the tricycle, and the vehicle speed at the time when the side bending occurs and other related situations. By extracting these data, it provides a detailed and crucial data basis for subsequent analysis.

[0031] For the accident road environment characteristics, similarity recognition is carried out for any two of the M pieces of data. Two pieces of data with a road environment similarity greater than a preset value are grouped into one category. In this way, N clustering clusters are gradually formed, where N must be less than M. Then, calculate the proportion of the data within each clustering cluster in the M pieces of braking accident data to obtain N proportion coefficients. At the same time, deeply analyze the distribution state of the braking accident characteristics corresponding to each clustering cluster, and establish N accident typical coefficients based on these distribution states. Finally, perform weighted processing on the N proportion coefficients and the N accident typical coefficients to obtain N scene value degrees. These scene value degrees and clustering clusters will provide an important basis for subsequent screening of effective accident road environment characteristics.

[0032] Define a preset value threshold, which is a pre-set standard used to measure the importance and reference value of the clustering cluster. Then, check the scene value degrees corresponding to the N clustering clusters one by one. For each clustering cluster, compare its scene value degree with the preset value threshold. If the scene value degree of a certain clustering cluster is greater than the preset value threshold, then this clustering cluster is considered to meet our screening conditions. Next, find out the effective accident road environment characteristics corresponding to these clustering clusters that meet the screening conditions. These characteristics are associated with the clustering clusters in the previous analysis process and contain various information about the road environment, such as the slope of the road, the road surface condition, the weather conditions, etc. Through such a screening process, select the effective accident road environment characteristics that are most helpful for establishing an effective test scene from among numerous clustering clusters and related characteristics, so as to improve the effectiveness and pertinence of the test scene and provide a more reliable basis for accurately evaluating the braking performance of electric tricycles in the future.

[0033] Configure the test scene carefully according to the selected effective accident road environment characteristics to establish an effective test scene. For the road slope in the road environment characteristics, set corresponding slope changes in the test site according to the slope data in the effective accident road environment characteristics, such as simulating different situations of steep slopes, gentle slopes, and flat ground. For the road surface condition, simulate the actual road surface according to the road surface smoothness and related information in the characteristics, such as using a sprinkler device to simulate a wet road surface and a special coating to simulate a smooth road surface. If weather characteristics are involved, use environmental simulation equipment to create corresponding weather conditions, such as a rainfall device for simulating rainy days and a snow-making device for simulating snowy days. At the same time, combined with the load constraint characteristics of the electric tricycle, place simulated loads in the front seat and cargo box according to the specified load requirements to fully restore the scene at the time of actual accidents. Through such meticulous configuration, establish an effective test scene that can accurately simulate the actual braking accident environment and provide a reliable environmental basis for subsequent braking tests.

[0034] In a possible implementation, step S420 further includes:

[0035] Step S421: Identify the similarity between any two of the M braking accident data based on the accident road environment characteristics, cluster any two data with a road environment similarity greater than the preset similarity, and generate the N clustering clusters.

[0036] Step S422: Calculate the proportion of the data within each of the N clustering clusters in the M braking accident data to generate N proportion coefficients.

[0037] Step S423: Analyze the distribution state of the braking accident characteristics corresponding to the N clustering clusters and establish N accident typical coefficients.

[0038] Step S424: Weight the N proportion coefficients and the N accident typical coefficients to generate the N scenario value degrees.

[0039] Specifically, identify the similarity between any two of the M braking accident data based on the accident road environment characteristics. For the road slope information in each piece of data, the ratio of the absolute value of the slope difference between the two pieces of data to the average slope can be calculated as the slope similarity index; for the road surface smoothness, by setting the values corresponding to different smoothness levels, the absolute value of the difference between the smoothness values corresponding to the two pieces of data can be calculated as the smoothness similarity index; for the weather characteristics, if the weather is the same, the similarity is 1, and if it is different, it is 0. Then, these three indexes are weighted and summed according to a certain weight (for example, the slope similarity weight is 0.4, the smoothness similarity weight is 0.3, and the weather similarity weight is 0.3) to obtain the comprehensive similarity. When the comprehensive similarity is greater than the preset similarity, these two pieces of data are clustered. Through pairwise comparison and clustering operations on all the data, N clustering clusters are finally generated.

[0040] For the already generated N clustering clusters, it is necessary to calculate the proportion of the data within each clustering cluster in the M braking accident data. For each clustering cluster, count the number of data contained therein, and then divide this number by the total number of the M braking accident data. The result obtained is the proportion coefficient corresponding to this clustering cluster. Through such a calculation method, a proportion coefficient is generated for each clustering cluster. These proportion coefficients can intuitively reflect the proportion of each clustering cluster in the entire data set, providing a quantitative basis for further analyzing the importance of the clustering clusters.

[0041] Conduct an in-depth analysis of the distribution status of braking accident characteristics corresponding to N clusters. For each cluster, carefully count the occurrence frequencies of various braking accident characteristics, such as the number of occurrences of situations like excessive braking distance, vehicle side bending, and braking failure. At the same time, considering the severity when these characteristics occur, it can be quantified by setting values corresponding to different severity levels. Then, based on this frequency and severity information, comprehensively calculate the characteristic values of each cluster, and then perform normalization processing on these characteristic values. Finally, establish N accident typical coefficients, which can accurately reflect the typical degree and importance of each cluster in terms of braking accident characteristics.

[0042] To generate N scenario value degrees, it is necessary to perform a weighting operation on N proportion coefficients and N accident typical coefficients. First, determine a reasonable weight allocation method. For example, different weight values can be assigned to the proportion coefficient and the accident typical coefficient according to the actual situation. Let the weight of the proportion coefficient be W1 and the weight of the accident typical coefficient be W2. Then for each cluster, multiply its corresponding proportion coefficient by W1, multiply the accident typical coefficient by W2, and then add the two products. The result obtained is the scenario value degree of this cluster. Through such a weighting calculation method, the proportion of the cluster in the dataset and the typicality of its braking accident characteristics are comprehensively considered, thereby generating N scenario value degrees that can comprehensively measure the comprehensive value of the cluster.

[0043] In a possible implementation manner, step S423 further includes:

[0044] Step S4231: Conduct a statistical analysis of the braking accident characteristics corresponding to the N clusters respectively, and establish N braking accident characteristic distribution states. Among them, any braking accident characteristic distribution state includes the proportional relationship of different types of braking accident characteristics.

[0045] Step S4232: Obtain a preset mapping relationship, where the preset mapping relationship includes the typical characteristic values corresponding to various braking accident characteristics, and the typical characteristic values are proportional to the accident severity.

[0046] Step S4233: After performing eigenvalue analysis on the N braking accident characteristic distribution states based on the preset mapping relationship, perform normalization processing to generate the N accident typical coefficients.

[0047] Specifically, for N clustering clusters, statistical analysis of braking accident characteristics is carried out one by one. For each clustering cluster, all the braking accident characteristics included in it are carefully sorted out, such as different types of characteristics like excessive braking distance, vehicle side bending, braking failure, and abnormal jitter during braking. By counting the frequency of each characteristic occurrence and calculating the proportion it occupies among all the braking accident characteristics in this clustering cluster, the braking accident characteristic distribution state corresponding to this clustering cluster is established. This distribution state is presented in the proportional relationship of different types of braking accident characteristics, clearly showing the composition of various braking accident characteristics in this clustering cluster and providing an important data basis for subsequent analysis.

[0048] It is necessary to obtain a preset mapping relationship. This preset mapping relationship is a pre-set rule set that covers the corresponding relationship between various braking accident characteristics and typical characteristic values. For each type of braking accident characteristic, there is a corresponding typical characteristic value. Moreover, this corresponding relationship is set according to the severity of the accident, that is, the higher the severity of the accident, the larger the corresponding typical characteristic value. For example, a severely excessive braking distance may correspond to a relatively high typical characteristic value, while a slight braking jitter may correspond to a relatively low typical characteristic value. By obtaining such a preset mapping relationship, it provides the necessary criteria and basis for subsequent eigenvalue analysis of the braking accident characteristic distribution state.

[0049] Carry out eigenvalue analysis based on the preset mapping relationship. For each braking accident characteristic distribution state, consider different types of braking accident characteristics among them. For each characteristic, calculate a comprehensive value according to its proportion in this distribution state and the corresponding typical characteristic value in the preset mapping relationship, that is, multiply the typical characteristic value of each characteristic by its proportion in this distribution state, and then add up all these products to obtain the comprehensive characteristic value of this distribution state. Next, perform normalization processing. Put the comprehensive characteristic values calculated for all braking accident characteristic distribution states together to form a set. For each comprehensive characteristic value in it, divide it by the maximum value in this set. After such processing, each comprehensive characteristic value is mapped to a value between 0 and 1, and this value is the accident typical coefficient corresponding to the clustering cluster finally obtained. In this way, an accident typical coefficient is generated for each clustering cluster, and these coefficients can accurately reflect the typical degree and importance of each clustering cluster in terms of braking accident characteristics.

[0050] In a possible implementation manner, step S500 further includes:

[0051] Step S510: Based on the effective test scenario, conduct a braking test on the electric tricycle to be detected, connect the sensing module to collect braking response data, and obtain a braking test data set.

[0052] Step S520: Extract the braking distance and braking response time based on the braking test data set to generate a braking timeliness test result.

[0053] Step S530: Conduct a braking force balance analysis based on the braking test data set to generate a braking force balance test result.

[0054] Step S540: Conduct a braking stability analysis based on the braking test data set to generate a braking stability test result.

[0055] Step S550: Generate the braking detection result using the braking timeliness test result, the braking force balance test result, and the braking stability test result.

[0056] Specifically, first, use an effective test scenario that simulates the characteristics of an actual accident road environment. Place the electric tricycle to be detected in it for braking tests, and at the same time connect professional sensing modules distributed at key parts such as wheels, braking systems, and vehicle bodies to accurately sense and record the changes in physical quantities during braking. When the vehicle brakes according to the predetermined process, the sensing modules synchronously and real-time collect braking response data such as wheel speed changes, braking pedal travel, braking force magnitude, and vehicle acceleration. After multiple test runs, summarize and organize these data to obtain a braking test data set for subsequent analysis of the vehicle's braking performance.

[0057] For the extraction of the braking distance, determine the data column related to the wheel speed in the data set, and find the data during the time period when the wheel speed drops from the pre-braking stable value to a complete stop. Based on the known parameter of the wheel circumference, through the integral operation of the product of the wheel speed and the time interval at each time point, the driving distance of the wheel during braking is obtained, which is the braking distance. For the extraction of the braking response time, locate the time point of the braking pedal trigger signal and the time point when the braking system pressure starts to rise significantly or the wheels start to show obvious deceleration in the data set, and subtract these two time points to obtain the braking response time. Then, compare the braking distance with the preset standard braking distance range and assign corresponding scores according to the deviation degree; do the same for the braking response time, and score it after comparing with the standard response time range. Finally, perform a weighted sum of these two scores according to a certain weight (such as a braking distance weight of 0.6 and a braking response time weight of 0.4), and the obtained result is the braking timeliness test result, which comprehensively and accurately measures the braking timeliness of the vehicle.

[0058] Deeply explore the braking test data set and accurately screen out the braking force data of each wheel during braking. For the analysis of the braking forces of the front and rear wheels, calculate the average braking forces of the front and rear wheels during the braking process in detail, and then find the ratio of the average braking forces of the front and rear wheels. Assuming that when the vehicle brakes normally, the ratio of the braking forces of the front and rear wheels should be within the reasonable range of 0.7 - 1.3. If the actually calculated ratio exceeds this range, it means that there is an imbalance in the braking forces of the front and rear wheels. Then analyze the braking forces of the left and right wheels, calculate the difference in the braking forces of the left and right wheels at each sampling moment, and record the maximum value of these differences. If the maximum value of the difference in the braking forces of the left and right wheels exceeds 10% of the total braking force of the vehicle, it can be determined that the braking forces of the left and right wheels are unbalanced. After completing the separate analysis of the braking forces of the front and rear, left and right wheels, combine the results of these two aspects and score according to the pre-set braking force balance evaluation system. For example, if both the ratio of the braking forces of the front and rear wheels and the difference in the braking forces of the left and right wheels are within the reasonable range, a full score is given; if one of them exceeds the range, corresponding deductions are made according to the degree of exceeding; if both do not meet the requirements, a lower score is given. Finally, generate the braking force balance test result based on the obtained score. This result can intuitively and accurately reflect the uniformity of the braking force when the vehicle brakes, providing a key basis for evaluating the stability and safety of the vehicle during braking, and thus clearly judging whether the vehicle will face potential risks such as loss of control, deviation, or side slip due to unbalanced braking forces.

[0059] Extract from the data set the longitudinal and lateral acceleration data of the vehicle during braking recorded by the acceleration sensor, as well as the change data of the pitch angle and roll angle of the vehicle monitored by the gyroscope. For the longitudinal acceleration, by analyzing its change curve from the start to the end of braking, judge whether there are sudden changes or abnormal fluctuations. If the acceleration changes smoothly, it indicates to a certain extent that the braking stability is good; on the contrary, if there are large sudden changes, it may imply that the braking process is unstable. At the same time, closely monitor the lateral acceleration data. If there is a large lateral acceleration value during braking and it exceeds the normal range, it means that the vehicle has a tendency to deviate or skid during braking, thus affecting the braking stability. Regarding the changes in the pitch angle and roll angle, observe the change amplitude during the braking process. If the angle changes slightly and can quickly return to the stable state, it shows that the vehicle attitude is maintained well; on the contrary, if the angle changes too much and lasts for a long time, it reflects that the braking has a greater impact on the stability of the vehicle body attitude. Combine the analysis results of these accelerations and angle changes, and quantify and score according to the pre-set braking stability evaluation criteria based on the deviation degree of each index. Finally, generate the braking stability test result to accurately determine the stable performance of the vehicle during braking, providing a key basis for the evaluation of vehicle braking safety.

[0060] Integrate the previously obtained braking timeliness test results, braking force balance test results, and braking stability test results organically to generate complete braking detection results. First, evaluate the braking timeliness test results to check whether the braking distance of the vehicle is within a reasonable range and whether the braking response time meets safety standards. If the braking distance is short and the braking response is rapid, a higher evaluation score can be given in this regard; otherwise, corresponding deductions are made according to the deviation degree from the standard value. Then analyze the braking force balance test results and score them based on the balance degree of the braking forces of the front and rear, left and right wheels. The more balanced the braking force, the higher the score; if there are obvious imbalances, the score is reduced because this will seriously affect the stability and safety of the vehicle during braking. Then consider the braking stability test results and quantitatively score them according to the stability degree of the pitch, roll, and lateral movement of the vehicle during braking. The better the stability performance, the higher the corresponding score. Finally, according to the predetermined weight distribution scheme, sum the scores of these three test results after weighting to obtain a comprehensive braking detection result. This result can comprehensively and accurately reflect the quality of the braking performance of the electric tricycle to be detected, provide a key basis for the safety assessment of the vehicle, and also provide a clear direction for possible subsequent improvement measures.

[0061] In a possible implementation manner, step S530 further includes:

[0062] Step S531: Extract the braking force distribution data of the front and rear, left and right wheels based on the braking test data set.

[0063] Step S532: Obtain the balanced braking force distribution characteristics of the front and rear, left and right wheels under the braking balance state.

[0064] Step S533: Compare the braking force distribution data with the balanced braking force distribution characteristics to determine the braking force offset vector and generate the braking force balance test result.

[0065] Specifically, conduct a comprehensive and meticulous review of the braking test dataset, and identify data records related to the braking forces of the front and rear wheels from the dataset. These records are sourced from information collected by devices such as pressure sensors and force sensors installed in the braking system. For the front-wheel braking force data, by identifying the data tags, filter out the braking force values generated by the front wheels at different moments during braking, and arrange them in chronological order to form the distribution data of the front-wheel braking force over time. Similarly, for the rear-wheel braking force, use a similar method to accurately extract the braking force values of the rear wheels at each time point from the numerous data, and then obtain the distribution of the rear-wheel braking force. When extracting the distribution data of the braking forces of the left and right wheels, based on the independent sensor data of the left and right wheels, such as torque data during wheel braking and friction data between the brake pads and brake discs, according to the same principle of extracting time series, sort out the braking force values of the left and right wheels respectively, so as to completely obtain the distribution data of the braking forces of the left and right wheels during the entire braking process, providing an accurate and detailed data basis for the subsequent evaluation of braking force balance.

[0066] By referring to relevant vehicle braking technical standards, theoretical models, and a large number of experimental data of similar vehicles in the ideal braking balance state, obtain the balanced braking force distribution characteristics that the front and rear, left and right wheels should possess in the braking balance state. These characteristics cover the reasonable proportional relationship between the braking forces of the front and rear wheels and the balance requirements of the braking forces of the left and right wheels, and are important bases for judging whether the current vehicle's braking force is balanced.

[0067] Compare the braking force distribution data of the front and rear wheels and left and right wheels extracted from the braking test dataset with the balanced braking force distribution characteristics obtained previously under the braking balance state. For the front and rear wheels, calculate the differences in the magnitude and action time of the actual braking force distribution data and the theoretical balanced braking force distribution characteristics, and obtain the braking force offset degree in the front-rear direction through difference calculation. For example, if the actual braking force of the front wheel is larger than the theoretical value at a certain moment, while the rear wheel is smaller, there will be an imbalance in the braking force in the front-rear direction. Determine the braking force offset vector component in the front-rear direction based on the magnitude and direction of this imbalance. Similarly, for the left and right wheels, compare their braking force distribution data with the balanced characteristics, consider the difference in the braking force of the left and right wheels and the time stage when the difference appears, determine the braking force offset degree in the left-right direction, and then obtain the braking force offset vector component in the left-right direction. Synthesize the braking force offset vector components in the front-rear and left-right directions to obtain a complete braking force offset vector, which comprehensively reflects the imbalance of the braking force. Finally, according to the pre-set braking force balance evaluation criteria, combined with the magnitude and direction of the braking force offset vector, quantitatively score the braking force balance degree of the vehicle to generate the braking force balance test result, so as to clearly present the balance performance of the braking force when the vehicle brakes and provide a key basis for the evaluation of the stability and safety of the vehicle braking system.

[0068] In a possible implementation manner, step S540 further includes:

[0069] Step S541: Extract the body motion posture data based on the braking test dataset, where the body motion posture data at least includes the pitch, roll, and lateral motion postures of the electric tricycle to be detected during braking.

[0070] Step S542: Perform posture stability analysis based on the body motion posture data to generate the braking stability test result.

[0071] Specifically, the long short-term memory network (LSTM), a machine learning algorithm, is used to extract the body motion posture data. First, the braking test data set is preprocessed, and the raw data collected from the acceleration sensor, gyroscope, and wheel speed sensor, etc. is normalized so that its numerical range is within a reasonable interval for the training and learning of the LSTM model. Then, the LSTM model architecture is constructed, which includes an input layer, a hidden layer, and an output layer. The input layer receives the preprocessed sensor data sequence, which contains various dynamic information of the vehicle during braking, such as the acceleration values, angular velocity values, and wheel speed values at different times, and is input into the model step by step according to the time steps. The hidden layer consists of multiple LSTM units, which can effectively capture the time series features and long-term dependencies in the data, and determine which information needs to be retained, forgotten, or updated through the internal gating mechanism, so as to model the motion state of the vehicle. In the model training stage, a large number of vehicle braking data samples with known motion postures are used to train the LSTM model. Through the backpropagation algorithm, the weights and bias parameters of the model are continuously adjusted to make the output of the model accurately predict the pitch, roll, and lateral motion postures of the vehicle as much as possible. For example, for the learning of the pitch motion posture, the model gradually masters the law of pitch angle change by analyzing the time series features of the front and rear axle acceleration changes and the data related to the body height in the input data; for the roll motion posture, it learns the rotation of the vehicle body around the longitudinal axis based on the angular velocity data sequence of the gyroscope; for the lateral motion posture, it combines the time series of the lateral acceleration and the wheel speed difference data to understand the lateral movement trend of the vehicle. When the model training is completed, the braking test data of the electric tricycle to be detected is input into the trained LSTM model. The model can output the corresponding body motion posture data according to the learned features and patterns, including accurate pitch angle values, roll angle values, and parameters such as lateral displacement, speed, and acceleration, thus completely extracting the body motion posture data of the vehicle during braking, providing a comprehensive and accurate data basis for the subsequent braking stability analysis, effectively improving the accuracy and efficiency of the extraction of vehicle motion posture data, and being able to better adapt to complex and changeable actual working conditions compared with traditional methods.

[0072] Analysis is carried out on the pitch attitude stability. According to the extracted vehicle body motion attitude data, the pitch angle change curve of the vehicle during braking is obtained. The change rate of the pitch angle is calculated by taking the first derivative of the pitch angle with respect to time. If the peak value of the change rate exceeds the preset safety threshold, it indicates that the vehicle has excessive nose diving during braking, which will affect the vehicle's controllability and stability. Because excessive pitch motion will cause large changes in the loads on the front and rear axles of the vehicle, resulting in uneven tire adhesion, and further increasing the risk of loss of control. At the same time, observe the recovery of the pitch angle after braking ends. If it cannot quickly recover to near the initial level, it also indicates that there are problems with the tuning of the vehicle's suspension system or braking system, affecting braking stability. Based on this, a score evaluation is made for the pitch attitude stability. Then analyze the roll attitude stability and check the roll angle change data of the vehicle. Statistically analyze the maximum value of the roll angle and its duration. If the maximum value of the roll angle exceeds the safety range allowed by the vehicle design, or if the roll angle remains large for a long time, it means that the vehicle may face the risk of rollover during braking, reflecting uneven braking forces on the left and right wheels or insufficient ability of the suspension system to resist roll, which will also have a serious negative impact on the braking stability of the vehicle. Based on these situations, corresponding scores are given for the roll attitude stability. For the lateral motion attitude stability, focus on the lateral displacement and lateral acceleration data of the vehicle. Calculate the standard deviation of the lateral displacement to measure the dispersion degree of the vehicle's lateral swing during braking. The larger the standard deviation, the more unstable the vehicle's driving trajectory. At the same time, check the peak value of the lateral acceleration. If it exceeds a certain limit, it may indicate that the vehicle is affected by a large lateral force interference, or the wheels are in an unstable state of alternating locking and sliding, which will reduce the braking stability of the vehicle. Quantitative scoring is performed according to the performance of each index of the lateral motion attitude. Finally, the score results of the pitch, roll, and lateral motion attitude stabilities are comprehensively weighted and summed according to the predetermined weight distribution scheme to generate the braking stability test result. This result can accurately reflect the overall stability performance of the vehicle during braking, provide a key basis for evaluating the braking safety of the vehicle, and help determine whether the vehicle can maintain a stable driving attitude during braking operations and avoid dangerous situations such as skidding and rollover.

[0073] In a possible implementation manner, step S300 further includes:

[0074] Step S310: Extract the historical load record data corresponding to any historical braking accident data from the historical braking accident data set.

[0075] Step S320: Based on the historical load record data, screen the historical braking accident data in the historical braking accident data set whose load characteristics meet the front seat load constraint and the cargo compartment load constraint to generate the effective braking accident data set.

[0076] Specifically, a dataset specifically storing historical braking accident data of electric tricycles is processed. This dataset is stored in a database or data storage system in a specific data format and structure, and its data sources cover multiple channels, such as accident records of traffic management departments, fault reports of vehicle repair factories, and claim data of relevant insurance institutions, etc., to ensure the comprehensiveness and diversity of the data. With the help of professional data extraction tools and technologies, the target historical braking accident data is quickly located through a data indexing mechanism. These data are presented in a structured form. For example, each piece of data contains multiple fields such as the accident occurrence time, location, vehicle driving state, braking system parameters, and load information unique to electric tricycles. Then, the record part related to the load is accurately separated from this piece of historical braking accident data, that is, the historical load record data. For electric tricycles, the historical load record data includes the weight information of passengers carried on the front seat (recorded or estimated through seat pressure sensors) and detailed data such as the weight, volume, and center of gravity position of the goods loaded in the cargo compartment (obtained through cargo compartment load sensors or estimation models based on the type and loading method of the goods). During the extraction process, data cleaning and preprocessing technologies are used to handle possible data missing, errors, or outliers. For example, if the front seat load data is missing, the average front seat load data of similar electric tricycles in similar accident scenarios can be referred to for reasonable filling; if the cargo compartment load data shows obvious anomalies, such as exceeding the design load limit of this vehicle type, it is verified and corrected in combination with the cargo list, transportation records, and the actual use of the vehicle, so as to ensure that the extracted historical load record data is accurate and provide a solid and reliable basis for subsequent screening of effective braking accident data based on load characteristics, to support the accurate evaluation and analysis of the braking performance of electric tricycles and ensure their safety and stability in actual operation.

[0077] Based on the obtained historical load record data, clarify the specific ranges of the front seat load constraint and the cargo compartment load constraint of the electric tricycle. The front seat load constraint is usually determined according to the designed load-bearing capacity of the vehicle, safety standards, and ergonomic principles. For example, through the test analysis of the seat structure strength and the vehicle handling stability under different passenger weights, a reasonable upper and lower limit range of the front seat passenger weight is set; the cargo compartment load constraint comprehensively considers factors such as the frame strength of the vehicle, the load-bearing capacity of the suspension system, and the performance of the power system under the condition of carrying goods, and determines the maximum safe weight of the cargo in the cargo compartment, the center of gravity position limit, and the reasonable range of the cargo distribution, etc. Then, traverse the entire historical braking accident data set, and compare the load characteristics in each data with the above-mentioned constraint conditions one by one. For the front seat load, check whether the recorded value is within the preset front seat load constraint range; for the cargo compartment load, not only judge whether the cargo weight meets the cargo compartment load constraint, but also evaluate whether the center of gravity position of the cargo and the uniformity of the cargo distribution meet the requirements. Through this meticulous screening process, select the historical braking accident data whose load characteristics fully meet the front seat load constraint and the cargo compartment load constraint, and integrate them to form a new data set, that is, the effective braking accident data set. This data set excludes the invalid data that may cause deviation in the results of braking accidents due to abnormal loads, making the subsequent braking performance analysis based on these data more scientific and accurate, and being able to more truly reflect the braking situation of the electric tricycle under normal load conditions, providing a reliable basis for evaluating and improving the braking system of the electric tricycle, thereby effectively improving the braking safety and stability of the vehicle during actual operation and reducing the risk of accidents caused by braking failures.

[0078] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the superiority or inferiority of the embodiments. And the above describes specific embodiments of this specification. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0079] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included within the protection scope of the present application.

[0080] This specification and the drawings are only exemplary descriptions of the present application and are considered to have covered any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. A brake detection method for an electric tricycle, characterized in that: include: Taking the type of electric tricycle to be detected as a constraint, a historical braking accident dataset is collected; Acquire the load constraint characteristics of the electric tricycle to be detected, wherein the load constraint characteristics include front seat load constraint and cargo compartment load constraint; Performing effective auxiliary data screening on the historical braking accident data set based on the front seat load constraint and the cargo compartment load constraint to generate an effective braking accident data set; configuring the road environment and load based on the effective braking accident data set to establish an effective test scenario; Based on the effective test scenario, a joint test of braking timeliness and braking stability is performed on the electric tricycle to be tested, and a braking test result of the electric tricycle to be tested is generated; The configuration of the road environment and load based on the effective braking accident data set to establish an effective test scenario includes: Extracting M pieces of braking accident data from the valid braking accident data set, where M is the total number of accidents corresponding to the valid braking accident data set, wherein any piece of braking accident data includes accident road environment characteristics and braking accident characteristics; Based on the accident road environment characteristics and the braking accident characteristics, the M braking accident data are clustered and the scene reference value is analyzed to obtain N clusters and N scene values, where N is a positive integer less than M; Based on the N clusters and the N scene values, screening effective accident road environment features corresponding to clusters whose scene values ​​are greater than a preset value threshold; Performing test scenario configuration based on the effective accident road environment characteristics to establish the effective test scenario; Wherein, clustering and scene reference value analysis are performed on the M pieces of braking accident data based on accident road environment characteristics and braking accident characteristics to obtain N clusters and N scene values, including: Based on the accident road environment characteristics, similarity identification is performed on any two pieces of the M brake accident data, and any two pieces of data whose road environment similarity is greater than a preset similarity are clustered to generate the N clusters; Calculating the proportion of the data in the clusters respectively corresponding to the N clusters in the M pieces of braking accident data, and generating N proportion coefficients; Analyze the distribution states of braking accident characteristics corresponding to the N clusters, and establish N typical accident coefficients; Weighting the N proportion coefficients and the N accident typical coefficients to generate the N scene values; The characteristic distribution states of braking accidents corresponding to the N clusters are analyzed to establish N typical accident coefficients, including: Performing statistical analysis on the braking accident features corresponding to the N clusters respectively, and establishing N braking accident feature distribution states, wherein any braking accident feature distribution state includes a proportional relationship of different types of braking accident features; Acquire a preset mapping relationship, wherein the preset mapping relationship includes typical characteristic values ​​corresponding to various types of braking accident characteristics, and the typical characteristic values ​​are proportional to the severity of the accident; Based on the preset mapping relationship, characteristic value analysis is performed on the N braking accident characteristic distribution states and then normalized to generate the N accident typical coefficients.

2. The brake detection method for an electric tricycle as claimed in claim 1, characterized in that: Based on the effective test scenario, a joint test of braking timeliness and braking stability is performed on the electric tricycle to be tested, and a braking test result of the electric tricycle to be tested is generated, including: Based on the effective test scenario, a braking test is performed on the electric tricycle to be tested, and a sensor module is connected to collect braking response data to obtain a braking test data set; Extracting the braking distance and the braking response time based on the braking test data set to generate a braking timeliness test result; Performing a braking force balance analysis based on the braking test data set to generate a braking force balance test result; Performing a braking stability analysis based on the braking test data set to generate a braking stability test result; The braking detection result is generated based on the braking effectiveness test result, the braking force balance test result and the braking stability test result.

3. A brake detection method for an electric tricycle as claimed in claim 2, characterized in that: Performing a braking force balance analysis based on the braking test data set to generate a braking force balance test result includes: Extracting braking force distribution data of front and rear wheels and left and right wheels based on the braking test data set; Obtaining the balanced braking force distribution characteristics of the front and rear wheels and the left and right wheels under the braking balance state; The braking force distribution data is compared with the balanced braking force distribution characteristics to determine the braking force offset vector and generate the braking force balance test result.

4. The brake detection method for an electric tricycle as claimed in claim 2, characterized in that: Performing a braking stability analysis based on the braking test data set to generate a braking stability test result includes: Extracting body motion posture data based on the braking test data set, wherein the body motion posture data at least includes the pitch, roll and lateral motion postures of the electric tricycle to be tested during the braking process; A posture stability analysis is performed based on the vehicle body motion posture data to generate the braking stability test result.

5. The brake detection method for an electric tricycle as claimed in claim 1, characterized in that: The effective auxiliary data of the historical braking accident data set is screened based on the front seat load constraint and the cargo compartment load constraint to generate an effective braking accident data set, including: Extracting historical load record data corresponding to any piece of historical braking accident data from the historical braking accident data set; Based on the historical load record data, the historical braking accident data whose load characteristics satisfy the front seat load constraint and the cargo compartment load constraint are screened in the historical braking accident data set to generate the effective braking accident data set.

Images