Dynamic prediction and risk assessment method for robot path planning and scheduling optimization

Abstract

The invention discloses a dynamic prediction and risk assessment method for robot path planning and scheduling optimization, and the method comprises the following steps: 1, environment data collection and processing; 2, event occurrence number calculation and standardization processing; step 3, integrating the event index into a path planning model; step 4, risk assessment and optimization; 5, fitting and verifying the model; and 6, performing dynamic adjustment and real-time optimization. The method not only makes significant progress in the field of robot path planning and scheduling optimization, but also has wide application prospects, and can play an important role in multiple fields of logistics, manufacturing, emergency rescue and the like.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence and robotics technology, and specifically to a dynamic prediction and risk assessment method for robot path planning and scheduling optimization. Background Art

[0002] At present, robot path planning and scheduling optimization technology mainly relies on the following common methods:

[0003] 1. Static path planning algorithm:

[0004] (1) A* algorithm: A* algorithm is a classic path planning algorithm that finds the shortest path from the starting point to the end point through heuristic search. Its advantage is that it is simple to calculate and suitable for static environments, but it performs poorly in dynamic environments.

[0005] (2) Dijkstra algorithm: Dijkstra algorithm finds the shortest path through breadth-first search. It is suitable for static environments, but its computational complexity is high and it is difficult to cope with large-scale environments.

[0006] 2. Dynamic path planning algorithm:

[0007] (1) D* algorithm: The D* algorithm is an extension of the A* algorithm and can replan paths in dynamic environments, but it has a large amount of calculations and is not efficient in complex environments.

[0008] (2) RRT (Rapid Random Tree) algorithm: The RRT algorithm generates paths through random sampling and is suitable for dynamic and unknown environments, but the path generation quality is unstable and requires a lot of computing resources.

[0009] 3. Methods based on probability and statistics:

[0010] (1) Markov decision process (MDP): MDP plans paths through a state transition probability model and is suitable for uncertain environments, but the model is complex and requires large amounts of computation.

[0011] (2) Particle filter: The particle filter estimates the path through statistical methods, but it has high computational complexity and poor adaptability.

[0012] Disadvantages of the prior art

[0013] Although the above methods have achieved certain results in path planning and scheduling optimization, they still have the following major defects in complex and dynamic environments:

[0014] (1) Poor adaptability to dynamic environment:

[0015] Most existing methods are based on the assumption of a static environment and have limited ability to handle dynamic changes in the environment. In practical applications, obstacles in the environment may move at any time, making static path planning methods ineffective.

[0016] (2) Lack of risk assessment mechanism:

[0017] Most existing path planning methods only consider the shortest path or the lowest cost path, and lack the assessment and control of path risks. This can easily lead to robots driving in high-risk areas, increasing the possibility of mission failure and damage.

[0018] (3) Single data processing and modeling:

[0019] Most existing methods have simple data processing and environment modeling, which makes it difficult to fully utilize the complex information in the environment. For example, the data and uncertainty processing of sensors are insufficient, which affects the accuracy of path planning.

[0020] (4) High computational complexity:

[0021] Although dynamic path planning and statistical methods can adapt to environmental changes to a certain extent, they have high computational complexity and poor real-time performance, making it difficult to meet the needs of real-time path planning and scheduling.

[0022] (5) Insufficient model scalability:

[0023] When dealing with multi-robot collaborative scheduling, existing methods have insufficient model scalability and are difficult to achieve efficient multi-robot collaborative work. The lack of an effective collaboration mechanism between robots can easily lead to resource waste and conflicts.

[0024] These defects limit the application of robots in complex and dynamic environments. There is an urgent need for a path planning and scheduling optimization method that can effectively respond to dynamic changes, has risk assessment capabilities, flexible data processing, high computational efficiency and good scalability.

[0025] Therefore, we propose a dynamic prediction and risk assessment method for robot path planning and scheduling optimization. Summary of the invention

[0026] The purpose of the present invention is to provide a dynamic prediction and risk assessment method for robot path planning and scheduling optimization, which solves the problems raised in the background technology.

[0027] To achieve the above-mentioned purpose, the present invention provides the following technical solution: a dynamic prediction and risk assessment method for robot path planning and scheduling optimization, comprising the following method steps:

[0028] Step 1: Environmental data collection and processing

[0029] Use sensors to collect environmental data in real time;

[0030] The collected data is preprocessed, including data cleaning, interpolation and correction, to ensure the integrity and accuracy of the data. The processing method is shown in the formula:

[0031] D clean =clean(D raw )

[0032] Among them, D clean is the cleaned data, D raw is the original data, clean(g) is the data cleaning operation;

[0033] Step 2: Calculation and standardization of event occurrences;

[0034] According to the environmental data, the number of events is calculated, and the calculated number of events is statistically processed and standardized to obtain the event index I t ,

[0035]

[0036] In the formula, I t represents the normalized event index, N t Indicates the number of events that occurred in the current time period, μ N and σ N are the mean and standard deviation of the number of events, respectively;

[0037] Step 3: Integrate event index into the path planning model

[0038] The event index is integrated into the path planning model for fitting. By dynamically adjusting the path planning strategy, the adaptability and accuracy of path planning are improved. The path planning model can be expressed as:

[0039]

[0040] Among them, P t represents the optimal path at time t, C(P) is the cost function of path P, R(P,I t ) is the path P at event index I t The risk under, λ is the risk weight coefficient;

[0041] Step 4: Risk Assessment and Optimization

[0042] The extended model CAViaR-E is used to calculate the path risk and evaluate the risk level of different paths. The extended model CAViaR-E of path risk is constructed, and the formula is as follows:

[0043] R t =β0+β1Rt-1 +β2I t +θ t

[0044] Among them, R t represents the path risk at time t, β0, β1, β2 are model parameters, and I t is the event index, θ t is the error term;

[0045] Step 5: Model fitting and verification

[0046] The path planning and scheduling optimization model is further fitted, and the likelihood ratio test method is used to evaluate the model fitting effect to ensure that the model has high reliability and accuracy and improve the robot's adaptability in complex and dynamic environments. The CAViaR-E model is used to fit the model, and the likelihood ratio test can be expressed as:

[0047] LR=-2(In L0-In L1)

[0048] Among them, LR is the likelihood ratio test statistic, L0 and L1 are the likelihood function values ​​of the original model and the new model respectively;

[0049] Step 6: Dynamic adjustment and real-time optimization

[0050] Finally, based on real-time environmental data and risk assessment results, the path planning and scheduling strategies are dynamically adjusted to ensure that the robot can cope with dynamic changes in the environment;

[0051] P t+1 =P t +α(R t +R t-1 )

[0052] Among them, P t+1 is the adjusted path planning result at time t+1, α is the adjustment coefficient, R t is the path risk at time t.

[0053] As a preferred embodiment of the present invention, the sensor in step one is, for example, a laser radar or a camera.

[0054] As a preferred implementation of the present invention, the number of event occurrences calculated in step 2 is specifically the frequency of obstacle occurrences.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] 1. Improve the path planning accuracy and scheduling optimization capabilities of robots in complex and dynamic environments.

[0057] Explanation: The present invention collects environmental data in real time and combines dynamic event prediction and risk assessment to adjust the path planning strategy in a timely manner, thereby improving the path planning accuracy and scheduling optimization capabilities of the robot in complex and dynamic environments.

[0058] Example: In a dynamically changing warehouse environment, the robot can adjust its path planning strategy by sensing obstacles and other dynamically changing factors in the environment in real time to ensure that the task is completed efficiently and safely.

[0059] 2. Reduce uncertainty and high-risk decisions in path planning and scheduling through dynamic event prediction and risk assessment.

[0060] Explanation: The event index and the extended model CAViaR-E are used to calculate the path risk, and the path selection is optimized based on the risk assessment results, effectively reducing the uncertainty and high-risk decisions in the path planning and scheduling process.

[0061] Example: In a motion task, by evaluating the risks of different paths and selecting the path with lower risks, the probability of unexpected situations during transportation can be reduced, thus improving the success rate of the task.

[0062] 3. Improve data processing and modeling methods to enhance the robot's ability to perceive and understand environmental information.

[0063] Interpretation: Ensure the integrity and accuracy of environmental data by cleaning, interpolating and correcting them, and improve the perception and understanding of environmental information by using standardized processing and event index calculation.

[0064] Example: In a mission, by preprocessing and standardizing sensor data, the location and movement of goods can be more accurately perceived, thereby optimizing robot path planning and scheduling strategies.

[0065] 4. It provides an efficient and reliable path planning and scheduling optimization solution, which is suitable for various complex and dynamic environments and improves the working efficiency and safety of robots.

[0066] Explanation: The present invention provides an efficient and reliable path planning and scheduling optimization method that can adapt to various complex and dynamic environments and improve the working efficiency and safety of the robot.

[0067] Example: In an emergency rescue scenario, the robot can sense environmental changes in real time, dynamically adjust the path and scheduling strategy, and complete the rescue mission quickly and safely.

[0068] 5. It has good scalability and adaptability, and can be applied to multi-robot collaborative work to improve the coordination and resource utilization of the overall system.

[0069] Explanation: The method of the present invention has good scalability and can be applied to the collaborative work of multiple robots to improve the coordination and resource utilization of the overall system.

[0070] Example: In a manufacturing workshop, multiple robots can work together to rationally allocate tasks and paths, improve production efficiency, and reduce resource waste and conflicts.

[0071] 6. Use advanced model fitting and verification methods to ensure high reliability and accuracy of the model.

[0072] Explanation: The extended model CAViaR-E was used for risk assessment, and the likelihood ratio test method was used to evaluate the model fitting effect to ensure the high reliability and accuracy of the model.

[0073] Example: During the path planning process, the model is continuously verified and optimized to ensure that the model can accurately reflect actual environmental changes and improve the reliability and safety of path planning.

[0074] Through the above-mentioned beneficial effects, the present invention has not only achieved remarkable progress in the field of robot path planning and scheduling optimization, but also has broad application prospects and can play an important role in logistics, manufacturing, emergency rescue and other fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0076] Figure 1 This is a flow chart of the dynamic prediction and risk assessment method for robot path planning and scheduling optimization of the present invention. DETAILED DESCRIPTION

[0077] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0078] The present invention provides a dynamic event prediction and risk assessment method for robot path planning and scheduling optimization, which specifically includes the following steps:

[0079] S1: Environmental data collection and processing

[0080] Use sensors (such as lidar, cameras, etc.) to collect environmental data in real time.

[0081] The collected data is preprocessed, including data cleaning, interpolation and correction, to ensure the integrity and accuracy of the data. The processing method is shown in the formula:

[0082] D clean =clean(Draw )

[0083] Among them, D clean is the cleaned data, D raw is the original data, and clean(g) is the data cleaning operation.

[0084] S2: Calculation and standardization of event occurrences

[0085] According to the environmental data, the number of events (such as the frequency of obstacles) is calculated, and the calculated number of events is statistically processed and standardized to obtain the event index I t ,

[0086]

[0087] In the formula, I t represents the normalized event index, N t Indicates the number of events that occurred in the current time period, μ N and σ N are the mean and standard deviation of the number of events, respectively.

[0088] S3: Integrating event index into path planning model

[0089] The event index is integrated into the path planning model for fitting. By dynamically adjusting the path planning strategy, the adaptability and accuracy of path planning are improved. The path planning model can be expressed as:

[0090]

[0091] Among them, P t represents the optimal path at time t, C(P) is the cost function of path P, R(P,I t ) is the path P at event index I t λ is the risk weight coefficient.

[0092] S4: Risk Assessment and Optimization

[0093] The extended model CAViaR-E is used to calculate the path risk and evaluate the risk level of different paths. The extended model CAViaR-E of path risk is constructed, and the formula is as follows:

[0094] R t =β0+β1R t-1 +β2I t +θ t

[0095] Among them, R t represents the path risk at time t, β0, β1, β2 are model parameters, and I t is the event index, θt is the error term.

[0096] Based on the risk assessment results, optimize the path planning and scheduling strategies and select the path and scheduling plan with the lowest risk.

[0097] S5: Model fitting and validation

[0098] The path planning and scheduling optimization model is further fitted, and the likelihood ratio test method is used to evaluate the model fitting effect to ensure that the model has high reliability and accuracy and improve the robot's adaptability in complex and dynamic environments. The CAViaR-E model is used to fit the model, and the likelihood ratio test can be expressed as:

[0099] LR=-2(In L0-In L1)

[0100] Among them, LR is the likelihood ratio test statistic, L0 and L1 are the likelihood function values ​​of the original model and the new model respectively.

[0101] S6: Dynamic adjustment and real-time optimization

[0102] Finally, based on real-time environmental data and risk assessment results, path planning and scheduling strategies are dynamically adjusted to ensure that the robot can cope with dynamic changes in the environment.

[0103] P t+1 =P t +α(R t +R t-1 )

[0104] Among them, P t+1 is the adjusted path planning result at time t+1, α is the adjustment coefficient, R t is the path risk at time t.

[0105] Embodiment 1

[0106] Warehouse robot path planning and scheduling

[0107] Environment Settings

[0108] Assume that there are multiple paths to choose from in the warehouse. The base cost C, number of events E, and normalized event index S of each path are known. We need to calculate the total cost of each path based on these data and select the path with the lowest cost.

[0109] Path data

[0110] Path 1: Basic cost C1 = 10, number of events E1 = 5

[0111] Path 2: Basic cost C2 = 12, number of events E2 = 3

[0112] Path 3: Basic cost C3 = 15, number of events E3 = 2

[0113] Step 1: Environmental data collection and processing

[0114] Collect environmental data and perform preprocessing to ensure data integrity and accuracy.

[0115] Step 2: Calculation and standardization of event occurrences

[0116] The event index is obtained by normalizing the number of events using a standardized formula.

[0117]

[0118] Assuming the mean μ = 3 and the standard deviation σ = 1, the event index of each path is calculated as follows:

[0119]

[0120] Step 3: Incorporate event index into the path planning model

[0121] Integrate the event index into the path planning model and dynamically adjust the path planning strategy. The path planning model formula is:

[0122] P i =C i +λS i

[0123] Here, it is assumed that λ=2.

[0124] Calculate the total cost of each path:

[0125] P1=10+2×2=14

[0126] P2=12+2×0=12

[0127] P3=15+2×(-1)=13

[0128] Step 4: Risk Assessment and Optimization

[0129] The extended model CAViaR-E is used to calculate the path risk. Assume that the path risk calculation formula is as follows:

[0130]

[0131] Among them, a=1, b=0.5, c=0.2.

[0132] Calculate the risk level of each path:

[0133] R1=1+0.5×2+0.2×2 2 =2.8

[0134] R2=1+0.5×0+0.2×0 2 =1

[0135] R3=1+0.5×(-1)+0.2×(-1) 2 =0.7

[0136] Step 5: Model fitting and validation

[0137] The likelihood ratio test method is used to fit and verify the path planning model to ensure its high reliability and accuracy. Assume that the likelihood value of the old model is L old =0.8, the likelihood value of the new model is L new =0.95,

[0138] L = 2 × (ln0.95-ln0.8) = 0.344

[0139] Step 6: Dynamic adjustment and real-time optimization

[0140] According to the real-time environmental data and risk assessment results, the path planning and scheduling strategies are dynamically adjusted. Assuming the real-time adjustment coefficient k = 0.1,

[0141] P i,调整 =P i +kR i

[0142] Calculate the final adjusted cost of each path:

[0143] P 1,调整 =14+0.1×0.28=14.28

[0144] P 2,调整 =12+0.1×1=12.1

[0145] P 3,调整 =13+0.1×0.07=13.07

[0146] Final path selection

[0147] According to the final adjusted path cost, select the path with the lowest cost. The final cost of each path is as follows:

[0148] Path 1: 14.28

[0149] Path 2: 12.1

[0150] Path 3: 13.07

[0151] The optimal path is path 2, and its final cost is 12.1.

[0152] Through these detailed numerical calculations and step-by-step demonstrations, the path planning model can effectively select the optimal path based on dynamic events and risk assessment, ensuring that the robot can complete the task efficiently and safely.

[0153] Embodiment 2

[0154] Path planning for urban logistics delivery robots

[0155] Environment Settings

[0156] Assume that there are multiple delivery routes available in the city. The base cost C, number of events E, and standardized event index S of each route are known. We need to calculate the total cost of each route based on these data and select the route with the lowest cost.

[0157] Path data

[0158] Path A: Base Cost C A =15, the number of events E A =4

[0159] Path B: Base Cost C B =18, the number of events E B =2

[0160] Path C: Base Cost C C =20, the number of events E C =5

[0161] Step 1: Environmental data collection and processing

[0162] Collect urban traffic environment data and perform preprocessing to ensure the integrity and accuracy of the data.

[0163] Step 2: Calculation and standardization of event occurrences

[0164] The event index is obtained by normalizing the number of events using a standardized formula.

[0165]

[0166] Assuming the mean μ = 3 and the standard deviation σ = 1, the event index of each path is calculated as follows:

[0167]

[0168] Step 3: Incorporate event index into the path planning model

[0169] Integrate the event index into the path planning model and dynamically adjust the path planning strategy. Path planning model formula:

[0170] P i =C i +λSi

[0171] Among them, λ=3.

[0172] Calculate the total cost of each path:

[0173] P A =15+3×1=18

[0174] P B =18+3×(-1)=15

[0175] P C =20+3×2=26

[0176] Step 4: Risk Assessment and Optimization

[0177] The path risk is calculated using the extended model CAViaR-E. According to the formula in the patent document, the path risk is calculated as follows:

[0178]

[0179] Among them, a=1, b=0.4, c=0.3.

[0180] Calculate the risk level of each path:

[0181] R A =1+0.4×1+0.3×1 2 =1.7

[0182] R B =1+0.4×(-1)+0.3×(-1) 2 =0.9

[0183] R C =1+0.4×2+0.3×2 2 =3

[0184] Step 5: Model fitting and validation

[0185] The likelihood ratio test method is used to fit and verify the path planning model to ensure its high reliability and accuracy. Assume that the likelihood value of the old model is L old =0.75, the likelihood value of the new model is L new =0.85,

[0186] L=2×(ln0.85-ln0.75)=0.2504

[0187] Step 6: Dynamic adjustment and real-time optimization

[0188] According to the real-time environmental data and risk assessment results, the path planning and scheduling strategies are dynamically adjusted. Assume that the real-time adjustment coefficient k = 0.1.

[0189] P i,调整 =P i +kR i

[0190] Calculate the final adjusted cost of each path:

[0191] P A,调整 =18+0.1×1.7=18.17

[0192] P B,调整 =15+0.1×0.9=15.09

[0193] P C,调整 =26+0.1×3=26.3

[0194] Final path selection

[0195] According to the final adjusted path cost, the path with the lowest cost is selected. The final cost of each path is as follows:

[0196] Path A: 18.17

[0197] Path B: 15.09

[0198] Path C: 26.3

[0199] The optimal path is path B, and its final cost is 15.09.

[0200] Through these detailed numerical calculations and step-by-step demonstrations, the path planning model can effectively select the optimal path based on dynamic events and risk assessment, ensuring that the robot can complete the task efficiently and safely.

[0201] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention.

[0202] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. A dynamic prediction and risk assessment method for robot path planning and scheduling optimization, characterized by: The method comprises the following steps: Step 1: Environmental data collection and processing; Use sensors to collect environmental data in real time; The collected data is preprocessed, including data cleaning, interpolation and correction, to ensure the integrity and accuracy of the data. The processing method is shown in the formula: D clean =clean(D raw ) Among them, D clean is the cleaned data, D raw is the original data, clean(g) is the data cleaning operation; Step 2: Calculation and standardization of event occurrences; According to the environmental data, the number of events is calculated, and the calculated number of events is statistically processed and standardized to obtain the event index I t , In the formula, I t represents the normalized event index, N t Indicates the number of events that occurred in the current time period, μ N and σ N are the mean and standard deviation of the number of events, respectively; Step 3: Integrate event index into path planning model; The event index is integrated into the path planning model for fitting. By dynamically adjusting the path planning strategy, the adaptability and accuracy of path planning are improved. The path planning model can be expressed as: Among them, P t represents the optimal path at time t, C(P) is the cost function of path P, R(P,I t ) is the path P at event index I t The risk under, λ is the risk weight coefficient; Step 4: Risk assessment and optimization; The extended model CAViaR-E is used to calculate the path risk and evaluate the risk level of different paths. The extended model CAViaR-E of path risk is constructed, and the formula is as follows: Among them, R t represents the path risk at time t, β0, β1, β2 are model parameters, and I t is the event index, is the error term; Step 5: Model fitting and verification The path planning and scheduling optimization model is further fitted, and the likelihood ratio test method is used to evaluate the model fitting effect to ensure that the model has high reliability and accuracy and improve the robot's adaptability in complex and dynamic environments. The CAViaR-E model is used to fit the model, and the likelihood ratio test can be expressed as: LR=-2(InL0-InL1) Among them, LR is the likelihood ratio test statistic, L0 and L1 are the likelihood function values ​​of the original model and the new model respectively; Step 6: Dynamic adjustment and real-time optimization Finally, based on real-time environmental data and risk assessment results, the path planning and scheduling strategies are dynamically adjusted to ensure that the robot can cope with dynamic changes in the environment; P t+1 =P t +α(R t +R t-1 ) Among them, P t+1 is the adjusted path planning result at time t+1, α is the adjustment coefficient, R t is the path risk at time t.

2. The dynamic prediction and risk assessment method for robot path planning and scheduling optimization according to claim 1, characterized in that: The sensor in step 1 is, for example, a laser radar or a camera.

3. The dynamic prediction and risk assessment method for robot path planning and scheduling optimization according to claim 1, characterized in that: The number of events calculated in step 2 is specifically the frequency of obstacle occurrence.

Images