Omnibearing intelligent anti-collision method and system for trackless equipment

Abstract

The invention discloses an omnibearing intelligent anti-collision method and system for trackless equipment, and belongs to the technical field of anti-collision, and the method comprises the steps: obtaining a preset motion track of the trackless equipment, dynamic obstacle state data and a static obstacle position; generating a dynamic obstacle prediction track according to the dynamic obstacle state data; establishing a risk analysis model according to the obstacle prediction trajectory, the static obstacle position and a preset motion trajectory of the trackless equipment, and generating a comprehensive risk value; according to the comprehensive risk value, establishing an optimal preset motion trail analysis model, and generating an optimal preset motion trail, so as to adjust the preset motion trail of the trackless equipment; according to the method, active prediction and accurate avoidance of obstacles in a complex dynamic environment are realized, the future motion trail of the dynamic obstacle can be predicted, comprehensive risk assessment of static and dynamic obstacles is realized, and the optimal motion trail is generated based on the risk cost, so that collision is effectively avoided, and the risk assessment efficiency is improved. And the operation continuity and the task execution efficiency of the trackless equipment are improved.

Description

Technical Field

[0001] This invention belongs to the field of intelligent anti-collision technology, and in particular relates to an all-round intelligent anti-collision method and system for trackless equipment. Background Technology

[0002] With the rapid development of intelligent manufacturing and logistics automation technologies, trackless equipment faces the dual challenges of static and dynamic obstacles when operating in complex and dynamic environments such as industrial production and warehousing logistics.

[0003] Existing technologies mainly rely on sensors to detect obstacles in real time and take emergency braking or fixed-path detour strategies, but lack the ability to predict the future trajectory of dynamic obstacles.

[0004] Specifically, traditional methods rely solely on current location information for reactive obstacle avoidance, failing to recognize the movement intentions of people or mobile devices, making it difficult to provide early warnings of potential collision risks. Furthermore, risk assessment mechanisms fail to distinguish between static and dynamic obstacles; static obstacle risk calculations only consider real-time distance, while dynamic obstacle risk assessments ignore trajectory prediction probability and time decay factors, resulting in insufficient accuracy in risk quantification models. In addition, path adjustment strategies are limited to single obstacle avoidance schemes, lacking an optimization selection mechanism based on multi-trajectory risk costs, leading to frequent device starts and stops or suboptimal detours, severely impacting operational continuity and task execution efficiency. These shortcomings make it difficult for trackless devices to achieve proactive, forward-looking, and comprehensive safety protection in scenarios with dense crowds or dynamically changing obstacles. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an all-around intelligent anti-collision method and system for trackless equipment, solving the aforementioned problems.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an all-around intelligent anti-collision method for trackless equipment, specifically comprising the following steps:

[0007] Acquire the preset motion trajectory, dynamic obstacle status data, and static obstacle positions of the trackless device;

[0008] Generate predicted trajectories for dynamic obstacles based on dynamic obstacle status data;

[0009] A risk analysis model is established based on the predicted obstacle trajectory, the static obstacle position, and the preset motion trajectory of the trackless equipment to generate a comprehensive risk value;

[0010] An optimal preset motion trajectory analysis model is established based on the comprehensive risk value, and the optimal preset motion trajectory is generated.

[0011] The preset motion trajectory of the trackless device is adjusted according to the optimal preset motion trajectory.

[0012] Based on the above technical solutions, the present invention also provides the following optional technical solutions:

[0013] Further technical solution: The method for generating the comprehensive risk value specifically includes the following steps:

[0014] Based on the location of static obstacles and the preset movement trajectory of the trackless equipment, a static obstacle risk value is generated;

[0015] Based on the predicted obstacle trajectory and the preset motion trajectory of the trackless device, a dynamic obstacle risk value is generated.

[0016] A risk analysis model is established based on the static and dynamic obstacle risk values ​​to generate a comprehensive risk value.

[0017] Further technical solution: The method for generating the static obstacle risk value specifically includes:

[0018] Through the formula:

[0019] ;

[0020] Generate static obstacle risk values ;

[0021] In the formula, This represents the static obstacle risk value of the m-th preset motion trajectory of the trackless device at time t. This represents the position of the trackless device at time t along the m-th preset motion trajectory. This represents the position of the i-th static obstacle. This represents the Euclidean distance between the trackless device and the i-th static obstacle at time t along the m-th preset motion trajectory. This represents the static obstacle distance warning value. For constant terms, This represents the risk intensity coefficient of the i-th static obstacle, and n represents the total number of static obstacles.

[0022] Further technical solutions: The specific method for generating the dynamic obstacle risk value includes:

[0023] Through the formula:

[0024] ;

[0025] Generate dynamic obstacle risk values ;

[0026] In the formula, This represents the dynamic obstacle risk value of the m-th preset motion trajectory of the trackless device at time t. This represents the probability that the predicted trajectory of the k-th dynamic obstacle will occur for the j-th dynamic obstacle. This represents the position of the trackless device at time t along the m-th preset motion trajectory. This represents the position of the k-th predicted trajectory of the j-th dynamic obstacle at time t. This represents the Euclidean distance between the position of the trackless device on the m-th preset trajectory at time t and the position of the j-th dynamic obstacle on the k-th predicted trajectory at time t. This represents the dynamic obstacle distance warning value. For constant terms, This represents the risk intensity coefficient of the j-th dynamic obstacle. This represents the time decay factor, and q represents the total number of dynamic obstacles. This represents the total number of predicted trajectories for the j-th dynamic obstacle;

[0027] The time decay factor is set to decrease from 1 to 0 as the time t increases from the current time, following a Gaussian curve.

[0028] Further technical solution: The specific expression of the risk analysis model is as follows:

[0029] ;

[0030] In the expression, This represents the comprehensive risk value of the m-th preset motion trajectory of the trackless device at time t. This represents the static obstacle risk value. This represents the dynamic obstacle risk value. , All are weighting coefficients, and .

[0031] Further technical solution: The method for generating the probability of the predicted trajectory of the dynamic obstacle specifically includes:

[0032] Through the formula:

[0033] ;

[0034] Generate the probability of dynamic obstacle prediction trajectory occurrence ;

[0035] In the formula, This represents the probability that the predicted trajectory of the k-th dynamic obstacle will occur for the j-th dynamic obstacle. This is the scaling factor. This represents the current direction vector of the j-th dynamic obstacle. This represents the direction vector of the k-th predicted trajectory of the j-th dynamic obstacle at time t. It represents the direction vector. With direction vector The size of the included angle between them This represents the direction vector of the x-th predicted trajectory of the j-th dynamic obstacle at time t. This represents the total number of predicted trajectories for the j-th dynamic obstacle.

[0036] Further technical solution: The method for generating the optimal preset motion trajectory specifically includes:

[0037] Based on the comprehensive risk value, generate the risk cost of trajectory replacement;

[0038] Based on the risk and cost of trajectory changes, an optimal preset motion trajectory analysis model is established to generate the optimal preset motion trajectory.

[0039] Further technical solution: The method for generating the trajectory replacement risk cost specifically includes:

[0040] Through the formula:

[0041] ;

[0042] Risk and cost of generating trajectory replacement ;

[0043] In the formula, This represents the overall risk value. This represents the overall risk threshold.

[0044] Further technical solution: The expression of the optimal preset motion trajectory analysis model is specifically as follows:

[0045] ;

[0046] In the expression, This represents the optimal preset motion trajectory of the trackless device. This represents the risk cost of trajectory replacement, and M represents the set of preset motion trajectories for the trackless device.

[0047] A trackless equipment omnidirectional intelligent collision avoidance system, the system being used to execute the aforementioned trackless equipment omnidirectional intelligent collision avoidance method, specifically including:

[0048] The data acquisition unit is used to acquire the preset motion trajectory, dynamic obstacle status data, and static obstacle position of the trackless device;

[0049] The trajectory prediction unit is used to generate predicted trajectories for dynamic obstacles based on dynamic obstacle state data.

[0050] The comprehensive risk analysis unit is used to establish a risk analysis model based on the predicted obstacle trajectory, the static obstacle position, and the preset motion trajectory of the trackless equipment, and generate a comprehensive risk value.

[0051] The optimal preset trajectory analysis unit is used to establish an optimal preset motion trajectory analysis model based on the comprehensive risk value and generate the optimal preset motion trajectory.

[0052] The control unit is used to adjust the preset motion trajectory of the trackless device according to the optimal preset motion trajectory;

[0053] The comprehensive risk analysis unit specifically includes:

[0054] The static obstacle analysis module is used to generate static obstacle risk values ​​based on the location of static obstacles and the preset movement trajectory of the trackless equipment.

[0055] The dynamic obstacle analysis module is used to generate dynamic obstacle risk values ​​based on the predicted obstacle trajectory and the preset motion trajectory of the trackless device.

[0056] The comprehensive risk value generation module is used to establish a risk analysis model based on static obstacle risk values ​​and dynamic obstacle risk values, and generate a comprehensive risk value.

[0057] The optimal preset trajectory analysis unit specifically includes:

[0058] The risk cost analysis module is used to generate the risk cost of trajectory replacement based on the comprehensive risk value.

[0059] The optimal preset motion trajectory output module is used to establish an optimal preset motion trajectory analysis model based on the risk cost of trajectory replacement and generate the optimal preset motion trajectory.

[0060] This invention provides an all-around intelligent anti-collision method and system for trackless equipment, which has the following advantages compared with the prior art:

[0061] This invention acquires data on the movement trajectory and obstacles of trackless equipment, generates a predicted trajectory for dynamic obstacles, establishes a comprehensive risk analysis model, and optimizes the movement trajectory based on risk cost. It achieves proactive prediction and precise avoidance of obstacles in complex dynamic environments. It has the advantages of being able to predict the future movement trajectory of dynamic obstacles, realize comprehensive risk assessment of static and dynamic obstacles, and generate the optimal movement trajectory based on risk cost, thereby effectively avoiding collisions and improving the continuity of trackless equipment operation and task execution efficiency. Attached Figure Description

[0062] Figure 1 This is a flowchart illustrating the all-around intelligent anti-collision method for trackless equipment provided by the present invention.

[0063] Figure 2 This is a flowchart illustrating step S30 of the present invention.

[0064] Figure 3 This is a flowchart illustrating step S40 of the present invention.

[0065] Figure 4 This is a flowchart illustrating step S20 of the present invention.

[0066] Figure 5 This is a structural schematic diagram of the trackless equipment all-around intelligent anti-collision system provided by the present invention.

[0067] Figure 6 This is a schematic diagram of the structure of the integrated risk analysis unit provided by the present invention.

[0068] Figure 7 This is a schematic diagram of the structure of the optimal preset trajectory analysis unit provided by the present invention. Detailed Implementation

[0069] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0070] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0071] Please see Figure 1 and Figure 4 The present invention provides an all-around intelligent anti-collision method for trackless equipment, comprising the following steps:

[0072] Step S10: Obtain the preset motion trajectory, dynamic obstacle status data, and static obstacle positions of the trackless device;

[0073] Step S20: Generate a predicted trajectory for the dynamic obstacle based on the dynamic obstacle state data;

[0074] Among them, the generation method of dynamic obstacle prediction trajectory can select the corresponding prediction model according to the category of dynamic obstacle; for example, for AGV / vehicle, a physical-intention hybrid model can be used; for personnel, a deep learning model (LSTM) can be used.

[0075] Step S30: Establish a risk analysis model based on the predicted obstacle trajectory, the static obstacle position, and the preset motion trajectory of the trackless equipment, and generate a comprehensive risk value;

[0076] Step S40: Establish an optimal preset motion trajectory analysis model based on the comprehensive risk value, and generate the optimal preset motion trajectory;

[0077] Step S50: Adjust the preset motion trajectory of the trackless device according to the optimal preset motion trajectory;

[0078] Among them, the generation method of dynamic obstacle prediction trajectory refers to the technical means of predicting the future movement path of dynamic obstacles based on dynamic obstacle state data. In practical applications, this generation method can select the corresponding prediction model according to the category of dynamic obstacle. For example, for AGV / vehicle, an extended Kalman filter can be used for trajectory prediction; for personnel, a variant of convolutional neural network combined with gated recurrent unit can be used for trajectory prediction. The main purpose is to realize the early recognition of the movement intention of dynamic obstacles.

[0079] Furthermore, the risk analysis model refers to a mathematical model used to quantify the collision risk between trackless equipment and obstacles. Specifically, this model can generate a comprehensive risk value by weighting the static obstacle risk value and the dynamic obstacle risk value; or it can generate a comprehensive risk value based on the risk level using a fuzzy logic system. Its main purpose is to accurately assess environmental risks.

[0080] The optimal preset trajectory analysis model can select the optimal path based on the risk and cost of trajectory change. Therefore, this embodiment constructs a closed-loop collision avoidance system by integrating environmental perception, trajectory prediction, and decision optimization mechanisms.

[0081] Specifically, by acquiring the preset motion trajectory of the trackless device, dynamic obstacle state data, and static obstacle positions as basic inputs, a complete characterization of the operating scenario is ensured. By generating a predicted dynamic obstacle trajectory, the limitation of traditional methods that rely solely on real-time position and cannot predict motion intentions is overcome. By establishing a risk analysis model to generate a comprehensive risk value, the distortion in risk assessment caused by considering only a single obstacle type in existing technologies is overcome. By generating and adjusting the optimal preset motion trajectory, the insufficient adaptability of traditional single-path planning in dynamic environments is avoided. Therefore, this application solves the problems of traditional collision avoidance methods in terms of missing dynamic obstacle prediction, coarse risk assessment, and insufficient trajectory optimization.

[0082] During the operation of the trackless equipment, preset motion trajectories, dynamic obstacle state data, and static obstacle positions are acquired to provide basic input for environmental perception, ensuring a complete depiction of the operating scenario. Specifically, predicted trajectories for dynamic obstacles are generated, with the generation method selecting the appropriate prediction model based on the type of dynamic obstacle. For example, a physical-intention hybrid model is used for AGVs or vehicles, while an LSTM deep learning model is applied for personnel, achieving high-precision trajectory prediction for the motion characteristics of different obstacles. Furthermore, the predicted obstacle trajectories, static obstacle positions, and preset motion trajectories are used to establish a risk analysis model, quantifying the comprehensive risk value. This model, by fusing predicted trajectories and static position information, accurately reflects the time-varying risk of dynamic obstacles and the fixed risk of static obstacles. Based on this, an optimal preset motion trajectory analysis model is established, the risk cost of trajectory replacement is calculated, and the optimal preset motion trajectory is generated. Finally, the preset motion trajectory is adjusted to respond to the optimal trajectory, ensuring the equipment adapts to environmental changes in real time.

[0083] In a specific embodiment, considering an intelligent warehousing and logistics scenario, the trackless equipment is an autonomous mobile robot. When dynamic obstacle status data shows that there are people walking ahead, the system selects an LSTM deep learning model based on the person's category to generate a prediction of their future movement trajectory. Simultaneously, when other AGVs are detected moving, a physical-intention hybrid model is used to predict their trajectories, and the position information of static obstacles such as fixed shelves is included in the analysis. A risk analysis model is constructed, a comprehensive risk value is calculated, and then the optimal movement path is determined through multi-trajectory scheme comparison to avoid potential conflicts with people, AGVs, and shelves.

[0084] Thus, the movement intention of dynamic obstacles can be identified in advance, overcoming the limitation of traditional methods that rely solely on real-time location and cannot predict movement intention; the establishment of a comprehensive risk model enables accurate quantification of the risks of static and dynamic obstacles, avoiding risk assessment distortion caused by a single obstacle type; the multi-track optimization mechanism provides a synergistic improvement in safety and operational efficiency, solving the problem of insufficient trajectory selection caused by a single obstacle avoidance strategy.

[0085] Specifically, in some of the embodiments described above in this application, a comprehensive risk value is proposed to assess the collision risk between trackless equipment and obstacles. However, in its implementation, it only mentions the establishment of a risk analysis model to generate a comprehensive risk value in a general way. It lacks a separate quantification mechanism and comprehensive method for the risks of static and dynamic obstacles, which makes it difficult for the risk assessment to accurately reflect the characteristic differences of different obstacle types and to achieve targeted risk control.

[0086] For this, please refer to Figure 2 The present invention further proposes a method for generating the comprehensive risk value, which specifically includes the following steps:

[0087] Step S31: Generate static obstacle risk values ​​based on the location of static obstacles and the preset motion trajectory of the trackless equipment;

[0088] Step S32: Generate dynamic obstacle risk values ​​based on the predicted obstacle trajectory and the preset motion trajectory of the trackless device;

[0089] Step S33: Establish a risk analysis model based on the static obstacle risk value and the dynamic obstacle risk value, and generate a comprehensive risk value;

[0090] Among them, the static obstacle risk value refers to the index that quantifies the collision risk between trackless equipment and static obstacles. It can be realized by using a calculation model based on the spatial distance attenuation characteristics, such as by evaluating the Euclidean distance relationship between the equipment position and the obstacle position. Its purpose is to independently handle the obstacle risk with a fixed position and avoid interference from dynamic factors.

[0091] The dynamic obstacle risk value can be understood as a risk assessment index that reflects the movement trend and uncertainty of dynamic obstacles. It can be implemented using a probability-weighted multi-trajectory prediction model. For example, different prediction algorithms can be selected based on the obstacle category and weighted according to the probability of trajectory occurrence. Its purpose is to proactively capture the potential collision threat of dynamic obstacles.

[0092] A risk analysis model specifically refers to a mathematical framework that integrates static and dynamic risk components. It can be implemented using a linear combination of adjustable weights, such as dynamically adapting weight coefficients to different scenario requirements. Its purpose is to generate a comprehensive indicator that can fully represent the overall risk level.

[0093] Specifically, the solution of this application first generates a static obstacle risk value based on the static obstacle position and preset motion trajectory in step S31. Since the static obstacle position is fixed, this step directly quantifies the risk based on the real-time spatial relationship between the equipment and the obstacle, ensuring that the assessment is not affected by dynamic interference. Based on this, step S32 generates a dynamic obstacle risk value based on the predicted obstacle trajectory rather than the current state. This step considers the movement trend and trajectory uncertainty of dynamic obstacles through a prediction model, giving the risk assessment a forward-looking perspective in the time dimension. Finally, step S33 integrates the static and dynamic obstacle risk values ​​through a risk analysis model. This model allows for dynamic adjustment of weighting coefficients according to the actual environment, thereby generating a comprehensive risk value that reflects the overall risk level. This step-by-step quantification and comprehensive assessment mechanism ensures that the risk analysis can accurately distinguish the characteristic differences between static and dynamic obstacles and provides a reliable basis for subsequent trajectory optimization.

[0094] As a specific implementation method, the solution of this application is implemented as follows: In an industrial warehousing scenario, the trackless equipment is an automated guided vehicle (AGV). Static obstacles include fixed shelves and wall structures, while dynamic obstacles include moving workers. The AGV first acquires shelf position data and calculates the static obstacle risk value based on a preset movement trajectory using a spatial distance attenuation model. Simultaneously, it monitors the movement status of workers through visual sensors, predicts their future walking trajectory using a deep learning model, and generates a dynamic obstacle risk value based on the probability of the trajectory occurring. Finally, the static obstacle risk value and the dynamic obstacle risk value are weighted and summed according to the weighting coefficients set according to the scenario requirements to obtain a comprehensive risk value, which is used to guide the AGV's path adjustment decisions.

[0095] Through the above scheme, this application achieves independent quantification and precise integration of the risks of static and dynamic obstacles, enabling risk assessment to effectively distinguish the physical characteristics and behavior patterns of different obstacle types. This supports immediate avoidance of static obstacles and predictive trajectory adjustment for dynamic obstacles, significantly improving the operational safety and path planning reliability of trackless equipment in complex dynamic environments.

[0096] Specifically, in some of the embodiments described above in this application, a method for generating static obstacle risk values ​​is proposed to achieve accurate quantification of the collision risk of static obstacles. However, in its implementation process, the existing method does not consider the risk intensity coefficient and distance sensitivity of the obstacle, which leads to the risk assessment relying solely on simple distance calculation and failing to distinguish the degree of danger of different static obstacles. Consequently, the risk value cannot accurately reflect the actual collision threat, affecting the effectiveness of subsequent trajectory optimization decisions.

[0097] In response, this invention further proposes a method for generating the static obstacle risk value, specifically including:

[0098] Through the formula:

[0099] ;

[0100] Generate static obstacle risk values ;

[0101] In the formula, This represents the static obstacle risk value of the m-th preset motion trajectory of the trackless device at time t. This represents the position of the trackless device at time t along the m-th preset motion trajectory. This represents the position of the i-th static obstacle. This represents the Euclidean distance between the trackless device and the i-th static obstacle at time t along the m-th preset motion trajectory. This represents the static obstacle distance warning value. For constant terms, This represents the risk intensity coefficient of the i-th static obstacle, and n represents the total number of static obstacles;

[0102] Among them, the risk intensity coefficient is a parameter used to quantify the inherent dangerous characteristics of static obstacles. It can be implemented by a linear mapping function based on the physical size of the obstacle or a discrete coding function based on the obstacle category. Its purpose is to enable risk assessment to distinguish between obstacles with different levels of danger, such as large equipment and small markers.

[0103] The static obstacle distance warning value is a reference threshold used to normalize the actual distance. It can be configured as the safe buffer distance required for braking of trackless equipment. Its purpose is to convert Euclidean distance into a relative distance ratio to reflect the nonlinear effects of distance changes.

[0104] The constant term is a small positive value used to ensure that the denominator is non-zero. Its purpose is to prevent the calculation process from diverging when the trackless device coincides with the position of the obstacle, and to maintain the numerical stability of the model.

[0105] Specifically, the proposed solution achieves dynamic quantification of static obstacle risk through the synergistic effect of a risk intensity coefficient and a distance normalization mechanism. The risk intensity coefficient is dynamically assigned based on obstacle attributes, giving higher weight to high-risk obstacles and thus addressing the problem of treating all obstacles equally in traditional methods. The distance normalization mechanism calculates the ratio between the actual Euclidean distance and the static obstacle distance warning value, causing the risk value to increase non-linearly as the distance decreases, with significantly enhanced sensitivity, especially when approaching the warning value, realistically simulating the characteristics of critical risk mutations. The introduction of a constant term ensures the model's computational robustness in extreme scenarios. Finally, the summation operation of all static obstacle risk contributions integrates all static risk sources in the environment, generating an overall static obstacle risk value that comprehensively reflects the risk level under the current preset motion trajectory.

[0106] As a specific embodiment, the solution of this application is implemented as follows: In an industrial warehousing environment, for static obstacles such as metal shelves, the risk intensity coefficient can be set to a higher value based on their physical size; for static obstacles such as ground marking lines, the risk intensity coefficient can be set to a lower value; the distance warning value of static obstacles is configured as the safe distance required for equipment braking; when the trackless equipment runs near the shelf, the risk value increases rapidly as the distance decreases, triggering the trajectory adjustment mechanism.

[0107] Through the above scheme, this application achieves a refined assessment of collision risk of static obstacles, which can effectively distinguish the degree of danger of different static obstacles, and make the risk value accurately represent the actual collision threat, thereby improving the accuracy and reliability of trajectory optimization decision-making for trackless equipment.

[0108] In some of the embodiments described above in this application, a method for generating dynamic obstacle risk values ​​is proposed to quantify the collision risk caused by dynamic obstacles. However, in its implementation, there is a lack of comprehensive consideration of the probability distribution of predicted trajectories of dynamic obstacles, the differences in risk intensity, and the influence of time factors. This results in the risk assessment relying solely on a single trajectory or static parameters, which cannot accurately reflect the uncertainty and time sensitivity of dynamic obstacle behavior. Consequently, the collision risk identification is inaccurate, affecting the effectiveness of obstacle avoidance decisions of trackless devices.

[0109] In response, this invention further proposes a method for generating the dynamic obstacle risk value, specifically including:

[0110] Through the formula:

[0111] ;

[0112] Generate dynamic obstacle risk values ;

[0113] In the formula, This represents the dynamic obstacle risk value of the m-th preset motion trajectory of the trackless device at time t. This represents the probability that the predicted trajectory of the k-th dynamic obstacle will occur for the j-th dynamic obstacle. This represents the position of the trackless device at time t along the m-th preset motion trajectory. This represents the position of the k-th predicted trajectory of the j-th dynamic obstacle at time t. This represents the Euclidean distance between the position of the trackless device on the m-th preset trajectory at time t and the position of the j-th dynamic obstacle on the k-th predicted trajectory at time t. This represents the dynamic obstacle distance warning value. For constant terms, This represents the risk intensity coefficient of the j-th dynamic obstacle. This represents the time decay factor, and q represents the total number of dynamic obstacles. This represents the total number of predicted trajectories for the j-th dynamic obstacle;

[0114] The risk intensity coefficient of dynamic obstacles can be determined by the size or type of the dynamic obstacles, and specifically by the flag function.

[0115] In practical applications, the probability of a predicted trajectory of a dynamic obstacle is a quantitative representation of the uncertainty of the dynamic obstacle's behavior. It can be achieved by using an exponential decay model based on the difference in the angle between the direction vectors or a statistical distribution model based on historical trajectory data. The purpose is to reflect the actual probability of different predicted trajectories and avoid the evaluation bias caused by relying solely on a single trajectory prediction.

[0116] The risk intensity coefficient can be understood as a differentiated parameter for the risk level of dynamic obstacles. It can be graded and assigned a value based on the physical size or type of the obstacle through a flag function. For example, a higher coefficient value can be assigned to large transport vehicles. The purpose is to make the risk assessment fit the actual differences in the severity of the collision consequences.

[0117] The time decay factor refers to the adjustment mechanism for time sensitivity in risk assessment. It can be implemented using an exponential decay function or a piecewise linear decay function. Its purpose is to prioritize high-probability collision events in the short term and enhance the timeliness of risk assessment.

[0118] In addition, the time decay factor can be selected in ways including but not limited to: decaying from 1 to 0 in a Gaussian curve as the time t is increasing from the current time.

[0119] The ratio of Euclidean distance to dynamic obstacle distance warning value is a normalized measure of spatial proximity. It is achieved by calculating the magnitude of the position vector difference and dividing it by a preset safety threshold. The purpose is to quantify real-time spatial risks while avoiding calculation anomalies where the denominator is zero.

[0120] The constant term is a small positive number that ensures the robustness of the calculation. It can be set to a fixed value within the precision range of floating-point numbers, with the aim of preventing the value from becoming unstable when the denominator approaches zero.

[0121] Specifically, the proposed solution first generates multiple predicted trajectories and their probabilities based on dynamic obstacle state data, with the probability calculated using the difference in the angle between direction vectors. Then, for each predicted trajectory, the Euclidean distance between the current position of the trackless device and the predicted position of the dynamic obstacle is calculated, and this distance is normalized to the dynamic obstacle distance warning value. Based on this, a risk intensity coefficient is used to weight the basic spatial risk; this coefficient is dynamically adjusted according to the obstacle type using a flag function. Furthermore, a time decay factor is introduced to dynamically correct the risk value, giving higher weight to recent high-probability events. Finally, a dual summation mechanism is used to probabilistically weight and accumulate the risk contributions of all dynamic obstacles and their predicted trajectories to form a comprehensive dynamic obstacle risk value. This process organically integrates the uncertainty of dynamic obstacle behavior, the differences in obstacle characteristics, and time sensitivity, ensuring that the risk assessment comprehensively reflects the true collision risk level in complex dynamic environments.

[0122] As a specific implementation method, the solution of this application is implemented as follows: In the intelligent warehousing and logistics scenario, when the system detects dynamic obstacles of personnel, the trajectory prediction unit uses an LSTM model to generate three possible predicted trajectories, one of which is a straight walking trajectory and the other two are turning trajectories; the risk intensity coefficient is set to a medium value according to the personnel's body shape characteristics through a flag function; the time decay factor is dynamically adjusted over time using an exponential decay function; the system calculates the Euclidean distance between the current position of the trackless equipment and each predicted trajectory point, and generates a normalized spatial risk by combining the dynamic obstacle distance warning value; then, the probability of occurrence of each trajectory is calculated based on the angle between the direction vectors, and the spatial risks are weighted and fused; finally, the output dynamic obstacle risk value is used by the comprehensive risk analysis unit to generate an accurate comprehensive risk assessment result.

[0123] Through the above technical solution, this application realizes refined modeling of dynamic obstacle risk, which can accurately identify the uncertainty and time sensitivity characteristics of dynamic obstacle behavior, effectively improve the accuracy of collision risk assessment, provide reliable obstacle avoidance decision basis for trackless equipment, and thus enhance the operational safety in complex dynamic environments.

[0124] In some of the embodiments described above in this application, a risk analysis model is proposed to generate a comprehensive risk value. However, in its implementation, the combination of static obstacle risk value and dynamic obstacle risk value lacks a clear quantitative mechanism and fails to fully consider the essential differences in risk characteristics between static obstacles (such as fixed structures) and dynamic obstacles (such as moving people or vehicles). This results in the risk assessment results failing to accurately reflect the contribution ratio of different obstacle types to the overall collision risk, thereby affecting the accuracy and safety of subsequent trajectory adjustments.

[0125] In response, this invention further proposes the following specific expression for the risk analysis model:

[0126] ;

[0127] In the expression, This represents the comprehensive risk value of the m-th preset motion trajectory of the trackless device at time t. This represents the static obstacle risk value. This represents the dynamic obstacle risk value. , All are weighting coefficients, and ;

[0128] Specifically, the comprehensive risk value refers to the overall collision risk quantification index of trackless equipment at a specific time and on a specific trajectory. It can be achieved by weighted summation, with the aim of comprehensively assessing the impact of different obstacle types on safe operation.

[0129] Weighting coefficient , It refers to the parameter used to adjust the proportion of static and dynamic risk contributions. It can be achieved by using scenario-adaptive adjustment, such as presetting different weight combinations according to the type of environment, with the aim of making the risk assessment closely fit the actual operating conditions.

[0130] Weighting coefficients and their constraints This can be understood as the mathematical condition that ensures the normalization of risk values. It can be achieved through normalization processing, with the aim of avoiding the distortion of assessment caused by weight imbalance and ensuring the reliability of risk quantification results.

[0131] Specifically, the solution in this application achieves refined quantification of the comprehensive risk value by linearly combining static and dynamic obstacle risk values ​​according to preset weights. The system first acquires the static and dynamic obstacle risk values, then dynamically adjusts the weighting coefficients based on the characteristics of the current operating environment. This increases the dynamic risk weight in areas with dense dynamic obstacles and increases the static risk weight in areas with dense static obstacles. Finally, a comprehensive risk value is generated through weighted summation. This design ensures that the risk assessment accurately reflects the essential differences in obstacle characteristics under different scenarios, providing a reliable decision-making basis for subsequent trajectory optimization, thus forming a complete risk quantification analysis chain.

[0132] As a specific embodiment, the solution of this application is implemented as follows: When the trackless equipment is running in a warehouse environment, the system detects a dense shelving structure and a small number of moving personnel in front. At this time, the risk analysis module automatically adjusts the weighting coefficients, making the risk weight of static obstacles higher than that of dynamic obstacles. The comprehensive risk value calculation unit receives the risk values ​​of static obstacles and dynamic obstacles, and generates a comprehensive risk value through a weighted summation formula. This value is then transmitted to the trajectory optimization module to generate a safe movement trajectory.

[0133] Through the above scheme, this application can dynamically adjust the risk weights of static and dynamic obstacles according to the actual operating environment, so that the comprehensive risk value accurately reflects the real risk situation in different scenarios, thereby improving the accuracy and safety of trajectory adjustment and effectively avoiding collision accidents caused by inaccurate risk assessment.

[0134] Specifically, in some of the embodiments described above in this application, the probability of the occurrence of a predicted trajectory of a dynamic obstacle is proposed to calculate the risk value of the dynamic obstacle. However, in its implementation, the generation of the probability lacks quantitative consideration of the consistency between the current movement direction of the dynamic obstacle and the direction of the predicted trajectory, which leads to the prediction of the trajectory probability being too general or relying on subjective assumptions. This fails to accurately reflect the actual movement intention of the dynamic obstacle, resulting in a large deviation in the calculation of the risk value of the dynamic obstacle and affecting the reliability of the overall collision avoidance decision.

[0135] In response, this invention further proposes a method for generating the probability of the predicted trajectory of the dynamic obstacle, specifically including:

[0136] Through the formula:

[0137] ;

[0138] Generate the probability of dynamic obstacle prediction trajectory occurrence ;

[0139] In the formula, This represents the probability that the predicted trajectory of the k-th dynamic obstacle will occur for the j-th dynamic obstacle. This is the scaling factor. This represents the current direction vector of the j-th dynamic obstacle. This represents the direction vector of the k-th predicted trajectory of the j-th dynamic obstacle at time t. It represents the direction vector. With direction vector The size of the included angle between them This represents the direction vector of the x-th predicted trajectory of the j-th dynamic obstacle at time t. This represents the total number of predicted trajectories for the j-th dynamic obstacle;

[0140] AngleDiff is a quantization function of the angle between direction vectors. It can be implemented by combining vector dot product with inverse cosine operation. For example, the cosine value of the angle is obtained by calculating the dot product of the normalized direction vectors and then converted into an angle. Its purpose is to transform the abstract difference in motion direction into a computable geometric parameter.

[0141] Scaling factor It can be understood as a direction sensitivity adjustment parameter, which can be set to a preset constant or dynamically configured according to the type of obstacle. For example, an adaptive adjustment mechanism can be used for obstacles to people to match their randomness of movement. Its purpose is to adapt to the differences in the movement characteristics of different dynamic obstacles.

[0142] This represents the size of the predicted trajectory set (i.e., the total number of predicted trajectories for dynamic obstacles). It can extract the main path branches from the output of multiple models through trajectory clustering algorithms, such as classifying motion intent categories based on historical trajectory data, with the aim of covering the potential motion possibilities of dynamic obstacles.

[0143] Specifically, the proposed solution constructs a probability generation logic through a directional consistency quantification mechanism. First, the current direction vector is obtained based on the real-time state of the dynamic obstacle, and the angle between this vector and the direction vector of each predicted trajectory at the corresponding moment is calculated. This angle directly reflects the degree of agreement between the predicted trajectory and the current movement trend. Second, the angle deviation is converted into a weight value through an exponential decay function; the smaller the angle, the greater the weight, reflecting the indicative role of directional consistency in movement intent. Subsequently, a scaling factor is used to adjust the sensitivity of the angle to the weights, enabling the model to distinguish the differences in movement characteristics between obstacles such as people and vehicles. Finally, the weights of all predicted trajectories are normalized and summed to form a probability distribution, ensuring mathematical rigor and comparability. This design closely links the real-time movement state of the dynamic obstacle with the predicted trajectory, avoiding the arbitrariness of the equal probability assumption and providing an objective basis for dynamic obstacle risk assessment.

[0144] As a specific implementation method, the solution of this application is implemented as follows: In the intelligent warehousing scenario, when the visual sensor detects a person as a dynamic obstacle, the system analyzes the position data of the person in continuous frames in real time and calculates the direction vector at the current moment; at the same time, multiple predicted trajectories are generated based on the LSTM model, and each trajectory corresponds to a specific direction vector at the target moment; subsequently, the angle between the current direction vector and the direction vector of each predicted trajectory is calculated and substituted into the above formula to generate a probability distribution; for example, if the angle between the direction of a certain predicted trajectory and the current direction is significantly smaller than that of other trajectories, its probability value is given a higher weight, and the system prioritizes the use of this trajectory for risk assessment.

[0145] Through the above technical solution, this application achieves accurate quantification of the probability of predicted trajectory of dynamic obstacles, effectively solves the problem of probability estimation deviation caused by insufficient consideration of directional consistency, thereby improving the accuracy of dynamic obstacle risk value calculation and enhancing the reliability of anti-collision decision of trackless equipment in complex dynamic environments.

[0146] In some of the embodiments described above in this application, an optimal preset motion trajectory analysis model is proposed to generate the optimal preset motion trajectory. However, in its implementation process, there is a lack of a quantification mechanism for the risk cost of trajectory replacement, which makes it impossible to accurately assess the risk differences of different preset motion trajectories, which may lead to the selection of high-risk trajectories or a reduction in equipment operating efficiency.

[0147] For this, please refer to Figure 3 The present invention further proposes a method for generating the optimal preset motion trajectory, specifically including:

[0148] Step S41: Generate the trajectory replacement risk cost based on the comprehensive risk value;

[0149] Step S42: Establish an optimal preset motion trajectory analysis model based on the risk cost of trajectory replacement, and generate the optimal preset motion trajectory;

[0150] Among them, trajectory change risk cost refers to an indicator used to quantify the risk change brought about by switching from the current preset motion trajectory to other trajectories. It can be implemented by means of the relative change rate of risk value, normalized representation of risk difference, or risk increment based on threshold. Its purpose is to transform the absolute risk value into a relatively comparable cost indicator, so as to facilitate objective decision-making in trajectory optimization.

[0151] The optimal preset motion trajectory analysis model can be understood as a decision-making mechanism for selecting the optimal trajectory from multiple preset motion trajectories. It can be implemented by using optimization algorithms that minimize cost functions, rule-based decision systems, or multi-objective optimization frameworks. Its purpose is to systematically compare the cost values ​​of different trajectory schemes and automatically select the preset motion trajectory with the lowest risk and highest efficiency.

[0152] Specifically, the solution in this application compares the comprehensive risk value with the comprehensive risk threshold in step S41 to generate a normalized trajectory replacement risk cost index, thereby quantifying and standardizing risk assessment and avoiding subjective judgments based solely on absolute risk values. Subsequently, in step S42, an analysis model is established based on these cost indices to systematically compare the cost values ​​of different preset motion trajectories and automatically select the trajectory with the lowest risk cost as the optimal preset motion trajectory. This two-step method ensures a close connection between risk assessment and trajectory optimization. By converting risk differences into quantifiable cost values, it provides an objective basis for trajectory optimization. At the same time, the core objective of minimizing risk costs ensures that the trajectory selection process considers both safety and operational efficiency, enabling trackless equipment to achieve more intelligent collision avoidance responses in dynamic environments.

[0153] As a specific implementation method, the solution of this application is implemented as follows: The processing unit can be implemented using an ARM Cortex-M7 microcontroller. After receiving the comprehensive risk value data from the risk analysis module, the microcontroller first compares these absolute risk values ​​with the comprehensive risk threshold using a preset algorithm to generate a normalized trajectory replacement risk cost index. Subsequently, based on these cost indices, the microcontroller uses an optimization algorithm that minimizes the cost function to select the trajectory with the lowest risk cost from multiple preset motion trajectories as the optimal preset motion trajectory, and sends this trajectory information to the control unit for execution. In actual operation, the microcontroller can process the risk data from the static obstacle analysis module and the dynamic obstacle analysis module in real time, ensuring the timeliness and accuracy of trajectory adjustment decisions.

[0154] Through the above solution, this application can accurately quantify the risk differences between different preset motion trajectories, avoid high-risk trajectory selection due to inaccurate risk assessment, and at the same time take into account the equipment operating efficiency, enabling trackless equipment to achieve a more intelligent and safer collision avoidance response in complex dynamic environments.

[0155] In some of the embodiments described above in this application, a trajectory replacement risk cost is proposed to quantify the risk level of different preset motion trajectories. However, in its implementation, due to the lack of a specific generation method, the calculation of risk cost lacks standardization and relativity, and cannot effectively reflect the difference between risk value and threshold, resulting in inaccurate selection of the optimal trajectory. It may misjudge the trajectory safety due to fluctuations in absolute risk value, affecting the operating efficiency and collision avoidance reliability of trackless equipment.

[0156] In response, this invention further proposes a method for generating the trajectory replacement risk cost, specifically including:

[0157] Through the formula:

[0158] ;

[0159] Risk and cost of generating trajectory replacement ;

[0160] In the formula, This represents the overall risk value. This represents the overall risk threshold;

[0161] Specifically, the risk cost of trajectory replacement refers to a relative indicator used to quantify the risk level of a preset motion trajectory. It can be calculated by the normalized difference between the comprehensive risk value and the comprehensive risk threshold. Its purpose is to transform the risk assessment from an absolute value to a relative degree of deviation, so as to avoid the distortion of comparison caused by the change of the risk value dimension.

[0162] The comprehensive risk value can be understood as a quantitative result that integrates the static obstacle risk value and the dynamic obstacle risk value. It can be obtained based on the output of the risk analysis model, and its purpose is to comprehensively reflect the comprehensive collision risk of environmental obstacles.

[0163] The comprehensive risk threshold can be understood as the system's preset safety boundary value, which can be determined by setting it based on experience or by dynamic adjustment. Its purpose is to provide a stable reference benchmark for risk assessment.

[0164] Specifically, the proposed solution generates a normalized trajectory replacement risk cost by comparing a comprehensive risk value with a preset risk threshold, thereby transforming risk assessment from an absolute value to a relative degree of deviation. When the comprehensive risk value is higher than the threshold, the risk cost is positive, directly reflecting the necessity of trajectory replacement; when it is lower than the threshold, the risk cost is negative, indicating that the trajectory is safe. The generated trajectory replacement risk cost is directly input into the optimal preset motion trajectory analysis model. By quantifying the relative magnitude of the risk costs of different trajectories, the system can accurately identify low-risk trajectories, avoiding the trajectory misselection problem caused by non-standardized risk calculation in traditional methods.

[0165] As a specific implementation method, the solution of this application is implemented as follows: The generation process of trajectory replacement risk cost is executed by a central processing unit. This processing unit receives the comprehensive risk value data stream from the comprehensive risk analysis unit and compares it with the pre-stored comprehensive risk threshold in real time. For example, the processing unit can use a floating-point arithmetic module to solve the formula and output a normalized risk cost value as the input parameter of the optimal preset motion trajectory analysis model. The calculation process can be implemented based on an embedded microcontroller to ensure the real-time performance and accuracy of the risk cost calculation.

[0166] Through the above technical solutions, the generation of trajectory replacement risk costs achieves the standardization and relativization of risk assessment, effectively solving the problem of the lack of relativity in risk cost calculation, avoiding trajectory misselection caused by fluctuations in absolute risk values, thereby improving the accuracy and operational efficiency of collision avoidance decisions for trackless equipment.

[0167] Specifically, in some of the embodiments described above in this application, an optimal preset motion trajectory analysis model is proposed to generate the optimal preset motion trajectory. However, in its implementation, the specific expression of the model is not clear, and it is impossible to quantify and evaluate the risk cost of each preset trajectory and select the optimal solution. This results in inaccurate obstacle avoidance decisions by the device in a dynamic environment, affecting operational efficiency and safety.

[0168] In response, this invention further proposes the following expression for the optimal preset motion trajectory analysis model:

[0169] ;

[0170] In the expression, This represents the optimal preset motion trajectory of the trackless device. This represents the risk cost of trajectory replacement, and M represents the set of preset motion trajectories for the trackless device.

[0171] Among them, the optimal preset motion trajectory refers to the optimal motion path selected by the trackless equipment during obstacle avoidance, with the aim of ensuring that the equipment maintains operating efficiency while avoiding obstacles.

[0172] The risk cost of trajectory replacement refers to a quantitative indicator for assessing the potential risks of each preset motion trajectory. It can be calculated based on the relative deviation between the comprehensive risk value and the risk threshold, with the aim of objectively comparing the safety of different trajectories.

[0173] The preset motion trajectory set M refers to a collection of multiple candidate motion trajectories pre-planned by the trackless device. These can include straight paths, curved paths, or combinations of paths, with the aim of providing diverse obstacle avoidance options to cope with complex environments.

[0174] Specifically, the solution in this application achieves automatic selection of the optimal preset motion trajectory by searching for the trajectory index that minimizes the trajectory replacement risk cost within a preset set of motion trajectories. The system first obtains the trajectory replacement risk cost corresponding to each trajectory in the preset set of motion trajectories M, which quantifies the potential risk level of each trajectory; subsequently, through... The operation is determined in M. The minimum trajectory index is used to output the optimal preset motion trajectory. This risk-cost minimization-based selection mechanism ensures the accuracy of obstacle avoidance decisions in dynamic environments and avoids the limitations of a single trajectory selection.

[0175] In one specific implementation, the preset motion trajectory set M of the trackless device can include three candidate paths: the first is a straight path, the second is a curved path that detours around the obstacle on the left, and the third is a curved path that detours around the obstacle on the right. The system calculates the trajectory replacement risk cost for each path. For example, the trajectory replacement risk cost of the first path is relatively high, the trajectory replacement risk cost of the second path is relatively low, and the trajectory replacement risk cost of the third path is moderate. The system determines that the path with the lowest trajectory replacement risk cost is the second path through a formula, and therefore selects the second path as the optimal preset motion trajectory.

[0176] Through the above scheme, this application can accurately quantify and assess the risk cost of each preset trajectory and select the optimal solution, thereby improving the obstacle avoidance decision accuracy of trackless equipment in dynamic environments and effectively enhancing operational efficiency and safety.

[0177] Please see Figure 5 In another embodiment, the present invention also proposes an all-around intelligent anti-collision system for trackless equipment, which is used to perform the above-described all-around intelligent anti-collision method for trackless equipment, specifically including:

[0178] Data acquisition unit 10 is used to acquire the preset motion trajectory, dynamic obstacle status data and static obstacle position of the trackless device;

[0179] The trajectory prediction unit 20 is used to generate a predicted trajectory for a dynamic obstacle based on the dynamic obstacle state data.

[0180] The comprehensive risk analysis unit 30 is used to establish a risk analysis model based on the predicted obstacle trajectory, the static obstacle position and the preset motion trajectory of the trackless equipment, and generate a comprehensive risk value.

[0181] The optimal preset trajectory analysis unit 40 is used to establish an optimal preset motion trajectory analysis model based on the comprehensive risk value and generate the optimal preset motion trajectory.

[0182] Control unit 50 is used to adjust the preset motion trajectory of the trackless device according to the optimal preset motion trajectory;

[0183] For this, please refer to Figure 6 The present invention further proposes that the comprehensive risk analysis unit 30 specifically includes:

[0184] The static obstacle analysis module 31 is used to generate static obstacle risk values ​​based on the location of static obstacles and the preset motion trajectory of the trackless equipment.

[0185] The dynamic obstacle analysis module 32 is used to generate dynamic obstacle risk values ​​based on the predicted obstacle trajectory and the preset motion trajectory of the trackless device.

[0186] The comprehensive risk value generation module 33 is used to establish a risk analysis model based on the static obstacle risk value and the dynamic obstacle risk value, and generate a comprehensive risk value;

[0187] For this, please refer to Figure 7 The present invention further proposes that the optimal preset trajectory analysis unit 40 specifically includes:

[0188] Risk cost analysis module 41 is used to generate trajectory replacement risk costs based on the comprehensive risk value;

[0189] The optimal preset motion trajectory output module 42 is used to establish an optimal preset motion trajectory analysis model based on the risk cost of trajectory replacement and generate the optimal preset motion trajectory.

[0190] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A omnidirectional intelligent anti-collision method for trackless equipment, characterized in that, The method specifically comprises the following steps: Obtaining a preset motion trajectory of the trackless equipment, dynamic obstacle state data, and static obstacle position; Generating a dynamic obstacle predicted trajectory according to the dynamic obstacle state data; Establishing a risk analysis model according to the obstacle predicted trajectory, the static obstacle position, and the preset motion trajectory of the trackless equipment, and generating a comprehensive risk value; Generating an optimal preset motion trajectory according to the comprehensive risk value; Adjusting the preset motion trajectory of the trackless equipment according to the optimal preset motion trajectory.

2. The omnibearing intelligent anti-collision method for trackless equipment according to claim 1, characterized in that, The generation mode of the comprehensive risk value specifically comprises the following steps: Generating a static obstacle risk value according to the static obstacle position and the preset motion trajectory of the trackless equipment; Generating a dynamic obstacle risk value according to the obstacle predicted trajectory and the preset motion trajectory of the trackless equipment; Establishing a risk analysis model according to the static obstacle risk value and the dynamic obstacle risk value, and generating a comprehensive risk value.

3. The omnibearing intelligent anti-collision method for trackless equipment according to claim 2, characterized in that, The generation mode of the static obstacle risk value specifically comprises: The generation mode of the dynamic obstacle risk value specifically comprises: ； Generating static obstacle risk values ; In the formula, represents the static obstacle risk value of the mth preset motion track of the trackless device at time t, represents the position of the trackless device in the mth preset motion track at time t, represents the position of the ith static obstacle, represents the Euclidean distance between the trackless device and the ith static obstacle in the mth preset motion track at time t, represents the static obstacle distance warning value, is a constant term, represents the risk intensity coefficient of the ith static obstacle, and n represents the total number of static obstacles.

4. The omnibearing intelligent anti-collision method for trackless equipment according to claim 2, characterized in that, The time attenuation factor is valued as: attenuating from 1 to 0 in a Gaussian curve form as the time length of the time t from the current time increases. The expression of the risk analysis model is specifically: ； Generating dynamic obstacle risk values ; In the formula, represents the dynamic obstacle risk value of the mth preset motion track of the trackless equipment at time t, represents the probability of the occurrence of the kth dynamic obstacle prediction track of the jth dynamic obstacle, represents the position of the trackless equipment in the mth preset motion track at time t, represents the position of the kth prediction track of the jth dynamic obstacle at time t, represents the Euclidean distance between the position of the trackless equipment in the mth preset motion track at time t and the position of the kth prediction track of the jth dynamic obstacle at time t, represents the dynamic obstacle distance warning value, is a constant term, represents the risk intensity coefficient of the jth dynamic obstacle, represents the time decay factor, and q represents the total number of dynamic obstacles, represents the total number of dynamic obstacle prediction tracks of the jth dynamic obstacle; The generation mode of the probability of the dynamic obstacle predicted trajectory occurring specifically comprises:

5. The omnidirectional intelligent anti-collision method for trackless equipment according to claim 2, characterized in that, The generation mode of the optimal preset motion trajectory specifically comprises: ； In the expression, represents the comprehensive risk value of the mth preset motion track of the trackless equipment at time t, represents the static obstacle risk value, represents the dynamic obstacle risk value, , are weight coefficients, and . 6.The omnibearing intelligent anti-collision method for trackless equipment according to claim 4, characterized in that, Generating a trajectory replacement risk cost according to the comprehensive risk value; Establishing an optimal preset motion trajectory analysis model according to the trajectory replacement risk cost, and generating an optimal preset motion trajectory. ； Generating a probability of occurrence of a dynamic obstacle predicted trajectory ; In the formula, represents the probability of the occurrence of the kth dynamic obstacle prediction trajectory of the jth dynamic obstacle, is a scaling coefficient, represents the current direction vector of the jth dynamic obstacle, represents the direction vector of the kth prediction trajectory of the jth dynamic obstacle at time t, represents the direction vector and the direction vector between them, represents the direction vector of the xth prediction trajectory of the jth dynamic obstacle at time t, represents the total number of dynamic obstacle prediction trajectories of the jth dynamic obstacle.

7. The omnibearing intelligent anti-collision method for trackless equipment according to claim 1, characterized in that, The generation mode of the trajectory replacement risk cost specifically comprises: The expression of the optimal preset motion trajectory analysis model is specifically: The system is used for executing the omnibearing intelligent anti-collision method of the trackless equipment according to any one of claims 1-9, and specifically comprises: 8.The omnibearing intelligent anti-collision method for trackless equipment according to claim 7, characterized in that, A data acquisition unit is configured to obtain a preset motion trajectory of the trackless equipment, dynamic obstacle state data, and static obstacle position; A trajectory prediction unit is configured to generate a dynamic obstacle predicted trajectory according to the dynamic obstacle state data; ； Generating trajectory change risk cost ; In the formula, represents a comprehensive risk value, represents a comprehensive risk threshold value. 9.The omnibearing intelligent anti-collision method for trackless equipment according to claim 7, characterized in that, A comprehensive risk analysis unit is configured to establish a risk analysis model according to the obstacle predicted trajectory, the static obstacle position, and the preset motion trajectory of the trackless equipment, and generate a comprehensive risk value; ； In the expression, represents the optimal preset motion trajectory of the trackless device, represents the trajectory replacement risk cost, and M represents the preset motion trajectory set of the trackless device.

10. A omnidirectional intelligent anti-collision system for trackless equipment, characterized in that, An optimal preset trajectory analysis unit is configured to establish an optimal preset motion trajectory analysis model according to the comprehensive risk value, and generate an optimal preset motion trajectory; A control unit is configured to adjust the preset motion trajectory of the trackless equipment according to the optimal preset motion trajectory. The comprehensive risk analysis unit specifically comprises: A static obstacle analysis module is configured to generate a static obstacle risk value according to the static obstacle position and the preset motion trajectory of the trackless equipment; A dynamic obstacle analysis module is configured to generate a dynamic obstacle risk value according to the obstacle predicted trajectory and the preset motion trajectory of the trackless equipment; A comprehensive risk value generation module is configured to establish a risk analysis model according to the static obstacle risk value and the dynamic obstacle risk value, and generate a comprehensive risk value; The optimal preset trajectory analysis unit specifically comprises: ​ ​ ​ ​ a risk cost analysis module, configured to generate a trajectory replacement risk cost according to the comprehensive risk value; an optimal preset motion trajectory output module, configured to establish an optimal preset motion trajectory analysis model according to the trajectory replacement risk cost, and generate an optimal preset motion trajectory.

Images