A mechanical driving collision detection system based on UWB positioning

Abstract

The invention discloses a mechanical driving collision detection system based on UWB positioning, and relates to the technical field of collision detection. The system comprises a positioning beacon, a UWB base station positioning system deployed in a construction area, which is responsible for receiving signals from the positioning beacon and calculating its position information, a construction intelligent safety supervision center, which monitors and manages the construction scope, the positioning beacon, the working scope of the staff and the operating machinery, and has an intelligent analysis function, an intelligent camera, and a monitoring platform. The invention adopts UWB technology for centimeter-level positioning to make the monitoring more accurate, adopts an intelligent analysis algorithm to quickly identify potential dangers, and realizes real-time early warning, adopts remote management and real-time alarm mechanism to improve management efficiency and emergency response speed; intruders, operating personnel, and supervisors jointly give alarms at multiple terminals to avoid accidents from multiple angles, thereby realizing comprehensive, accurate, and rapid monitoring and management of personnel and machinery in the construction area, and effectively preventing the occurrence of collision accidents.

Description

Technical Field

[0001] The present invention relates to the technical field of collision detection, and in particular to a mechanical driving collision detection system based on UWB positioning. Background Art

[0002] Today's construction site environment is very complex, filled with various workers and various construction machinery. Due to the complex construction site environment and real-time changes in location, the limited operating field of mechanical equipment leads to frequent collisions between personnel and mechanical equipment, posing a serious threat to the life safety of construction workers. In response to the above problems, some construction sites have formulated safety management methods, including process specifications, setting up cordons, wearing reflective clothing, manual monitoring, etc. However, these methods have the disadvantages of great management difficulty, poor real-time performance, and low reliability, making it difficult to effectively avoid collisions between personnel and mechanical equipment.

[0003] With the development of Internet of Things technology, some construction sites have begun to adopt positioning systems based on GPS or RFID. However, these technologies have problems such as insufficient positioning accuracy and susceptibility to environmental interference, making it difficult to meet the needs of high-precision, real-time safety supervision. At the same time, anti-collision technology based on visual recognition has disadvantages such as high computing and recognition delays and poor real-time performance. Therefore, achieving smarter, more specific environment-applicable, and more real-time intelligent mechanical anti-collision to ensure intelligent anti-collision throughout the entire process has become a key requirement for improvement. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a mechanical driving collision detection system based on UWB positioning to solve the problems of being unable to accurately distinguish the specific locations of construction personnel and the operating range of machinery, which easily leads to false alarms or missed alarms; the time from detecting danger to issuing an alarm is long, and the collision cannot be prevented in time; there is a lack of intelligent analysis capabilities and remote management means, and it is difficult to adapt to the complex and changeable construction environment.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] On the one hand, the present invention provides a mechanical driving collision detection system based on UWB positioning, which comprises:

[0008] Positioning beacons are attached to workers including construction personnel and drivers of construction vehicles, and to operating machinery including construction machinery and equipment and construction vehicles;

[0009] The UWB base station positioning system is deployed in the construction area and is responsible for receiving signals from positioning beacons and calculating their location information;

[0010] The construction intelligent safety supervision center monitors and manages the construction scope, positioning beacons, and the operating scope of workers and operating machinery, and has intelligent analysis functions;

[0011] Smart camera to monitor helmet wearing conditions;

[0012] The monitoring platform displays the real-time location information of the positioning beacon, receives and displays alarm information, and is equipped with a buzzer device to support remote speaking function.

[0013] On the other hand, the present invention provides a mechanical driving collision detection method based on UWB positioning, comprising:

[0014] Step S1, positioning and data transmission, the UWB base station positioning system receives the signal of the positioning beacon, calculates and transmits the location information to the construction intelligent safety supervision center;

[0015] Step S2, intelligent analysis, the construction intelligent safety supervision center determines whether to enter the danger zone based on the location information and the distance between the construction personnel and the operating machinery;

[0016] Step S3, real-time warning: once a potential collision risk is detected in step S2, an alarm message is immediately issued through the monitoring platform, and a real-time alarm is issued through a buzzer device;

[0017] Step S4, remote management, the administrator of the monitoring platform views the location information in real time, remotely calls out to the violators, and corrects the unsafe behavior in time.

[0018] Furthermore, in step S1, after receiving the signal of the positioning beacon, the UWB base station positioning system determines the position of the beacon and transmits the position information to the construction intelligent safety supervision center. The positioning beacon position is determined in the following manner:

[0019] The positioning beacon sends a signal to each base station. Assuming that n UWB base stations are deployed in the construction area, the propagation time t of the signal from the beacon to the i-th base station is i Expressed as Among them, t i represents the signal propagation time from the beacon to the i-th base station, d i represents the distance from the beacon to the i-th base station, and c represents the speed of light;

[0020] Calculate the time difference between the beacon arrival at different base stations. For base station i and base station j, the arrival time difference of the signal is Δt ij for:

[0021] Among them, Δt ij represents the time difference between the beacon reaching base station i and base station j, d j represents the distance from the beacon to the jth base station,

[0022] Assume that the position of the beacon is (x, y, z) and the position of base station i is (x i ,y i ,z i ), the location of base station j is (x j ,y j ,z j ), then the distance difference between the two base stations is Δd ij for:

[0023] Among them, (x i ,y i ,z i ) represents the known position coordinates of the i-th base station, (x j ,y j ,z j ) represents the known location coordinates of the jth base station.

[0024] Further, in step S1, the method for determining the location of the positioning beacon also includes:

[0025] For Δd ij Perform Taylor expansion and ignore higher-order terms to obtain the linearized equation:

[0026] The approximate linear equation obtained by expansion;

[0027] The linearized equations of multiple base stations are organized into matrix form, and the least squares method is used to solve the position of the beacon:

[0028] AX=B,where Among them, A represents the coefficient matrix formed by the position differences of each base station, X represents the unknown position vector of the beacon, and B represents the vector formed by the product of the time difference and the speed of light.

[0029] Furthermore, in step S1, the method for determining the position of the positioning beacon also includes: calculating X, that is, the three-dimensional coordinates (x, y, z) of the beacon by the least squares method, and the position information (x, y, z) and the timestamp t together constitute a data packet, and transmitting the data packet to the construction intelligent safety supervision center through the network.

[0030] Further, in step S2, the intelligent analysis method is:

[0031] The construction intelligent safety supervision center analyzes the construction site based on the location information in step S1 through the deep learning model CNN and the dynamic risk assessment model MDP;

[0032] Analyze the relative position and motion status between construction personnel and operating machinery based on position information.

[0033] Further, in step S2, the intelligent analysis method includes:

[0034] From step S1, we get the position information P = (x, y, z) and timestamp t. For beacon i, we transform its state vector s i Defined as: Among them, s i (t) represents the state vector of beacon i at time t, x i (t),y i (t),z i (t) represents the coordinates of beacon i at time t, represents the velocity component of beacon i at time t;

[0035] Construct the time series input. For the time window T, the time series input tensor S of beacon i i Expressed as: S i =[s i (tT),s i (t-T+1),…,s i (t)], where S i represents the time series input tensor of beacon i, and T represents the time window size.

[0036] Furthermore, the intelligent analysis step in step S2 further includes:

[0037] Use a one-dimensional convolutional neural network to process time series data and extract feature representation. For the input sequence S i , the convolution operation is defined as:

[0038] in, represents the feature representation of beacon i after the lth layer of convolution, represents the weight matrix of the convolution kernel of the lth layer, represents the feature representation of beacon i at layer l-1, b (l) represents the convolution bias of the lth layer, σ(·) represents the nonlinear activation function ReLU, K represents the convolution kernel length, and k represents the index variable in the convolution operation;

[0039] Perform pooling operation: in, It represents the feature representation of beacon i after pooling at layer l. After multiple layers of convolution and pooling operations, the feature vector f of beacon i is finally obtained. i : where f i represents the feature vector of beacon i, and Flatten means flattening the high-dimensional feature representation into a one-dimensional vector.

[0040] Furthermore, the dynamic risk assessment method in step S2 is:

[0041] Define the state space S in the MDP, where each state consists of the feature vectors of all current beacons:

[0042] s(t)=[f1,f2,…,f n ], where s(t) represents the state of the system at time t, and n represents the number of beacons;

[0043] Define the action space A, where actions represent the possible actions that the system may take when a risk is detected;

[0044] The state transition probability matrix P(s′|s,a) is defined to represent the probability that the system will transfer to the next state s′ after executing action a in the current state s, which is defined as: Where P(s′|s,a) represents the transition probability from state s to s′, T(s,a) represents the change function of state s caused by action a, and Z represents the normalization constant;

[0045] Define the reward function R(s,a), which represents the immediate reward for performing action a in state s:

[0046] Among them, γ i represents the weight of beacon i, n represents the number of beacons i, g i (a) represents the expected feature vector of beacon i after executing action a, and defines the state value function V(s), which represents the maximum expected reward obtained in state s: Among them, V(s) represents the value function at state s, ∑ s′ (·) represents the sum of all possible next states;

[0047] The reinforcement learning method QLearning is used to optimize the strategy π(s) so that the action a selected in each state s maximizes the long-term reward:

[0048] Among them, π(s) represents the optimal strategy in state s.

[0049] Furthermore, in step S2, the risk in each state is graded according to the optimized strategy π(s), and mapped to different risk levels through the value function V(s) and transmitted to the monitoring platform;

[0050] When step S2 detects a potential collision risk, the construction intelligent safety supervision center immediately sends out an alarm message through the monitoring platform.

[0051] The beneficial effects of the present invention are:

[0052] The present invention adopts UWB technology to perform centimeter-level positioning, so that the monitoring accuracy is higher.

[0053] The present invention adopts an intelligent analysis algorithm to quickly identify potential dangers and implement real-time early warning.

[0054] The present invention has a remote management and real-time alarm mechanism to improve management efficiency and emergency response speed; intruders, operators and supervisors can jointly give alarms at multiple terminals to avoid accidents from multiple angles.

[0055] In summary, the present invention realizes comprehensive, accurate and rapid monitoring and management of personnel and machinery in the construction area through high-precision positioning, intelligent analysis and real-time early warning, and effectively prevents the occurrence of collision accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0057] Figure 1 It is a schematic diagram of the process flow of the mechanical driving collision detection method of the present invention;

[0058] Figure 2 It is a schematic diagram of the structure of the mechanical driving collision detection system of the present invention. DETAILED DESCRIPTION

[0059] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0060] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0061] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0062] Example 1, reference Figure 2 , this embodiment provides a mechanical driving collision detection system based on UWB positioning, including:

[0063] Positioning beacons are attached to workers including construction personnel and drivers of construction vehicles, and to operating machinery including construction machinery and equipment and construction vehicles;

[0064] The UWB base station positioning system is deployed in the construction area and is responsible for receiving signals from positioning beacons and calculating their location information;

[0065] The construction intelligent safety supervision center monitors and manages the construction scope, positioning beacons, and the operating scope of workers and operating machinery, and has intelligent analysis functions;

[0066] Smart camera to monitor helmet wearing conditions;

[0067] The monitoring platform displays the real-time location information of the positioning beacon, receives and displays alarm information, and is equipped with a buzzer device to support remote speaking function.

[0068] Example 1, reference Figure 1 This embodiment provides a mechanical driving collision detection method based on UWB positioning, comprising the following steps:

[0069] Step S1, positioning and data transmission, the UWB base station positioning system receives the signal of the positioning beacon, calculates and transmits the location information to the construction intelligent safety supervision center;

[0070] In step S1, after receiving the signal of the positioning beacon, the UWB base station positioning system determines the position of the beacon and transmits the position information to the construction intelligent safety supervision center. The positioning beacon position is determined in the following way:

[0071] The positioning beacon sends a signal to each base station. Assuming that n UWB base stations are deployed in the construction area, the propagation time t of the signal from the beacon to the i-th base station is i Expressed as Among them, t i represents the signal propagation time from the beacon to the i-th base station, d i represents the distance from the beacon to the i-th base station, and c represents the speed of light;

[0072] Calculate the time difference between the beacon arrival at different base stations. For base station i and base station j, the arrival time difference of the signal is Δt ij for:

[0073] Among them, Δt ij represents the time difference between the beacon reaching base station i and base station j, d j represents the distance from the beacon to the jth base station,

[0074] Assume that the position of the beacon is (x, y, z) and the position of base station i is (x i ,yi ,z i ), the location of base station j is (x j ,y j ,z j ), then the distance difference between the two base stations is Δd ij for:

[0075] Among them, (x i ,y i ,z i ) represents the known position coordinates of the ith base station, (x j ,y j ,z j ) represents the known position coordinates of the j-th base station;

[0076] In step S1, the method for determining the location of the positioning beacon also includes:

[0077] For Δd ij Perform Taylor expansion and ignore higher-order terms to obtain the linearized equation:

[0078] The approximate linear equation obtained by expansion;

[0079] The linearized equations of multiple base stations are organized into matrix form, and the least squares method is used to solve the position of the beacon:

[0080] AX=B,where Where A represents the coefficient matrix formed by the position differences of each base station, X represents the unknown position vector of the beacon, and B represents the vector formed by the product of the time difference and the speed of light;

[0081] In step S1, the method for determining the position of the positioning beacon also includes: calculating X, that is, the three-dimensional coordinates (x, y, z) of the beacon by the least squares method, the position information (x, y, z) and the timestamp t together constitute a data packet, and transmitting the data packet to the construction intelligent safety supervision center through the network;

[0082] Specifically, the UWB base station positioning system accurately calculates the position of beacons within the construction site and transmits the location information to the construction intelligent safety supervision center; the UWB base station positioning system realizes high-precision positioning of the three-dimensional spatial position of each beacon (including personnel and mechanical equipment) within the construction site. UWB technology measures the signal propagation time difference between the beacon and the base station, and combines the time difference data of multiple base stations to accurately calculate the three-dimensional coordinates of the beacon, providing extremely high positioning accuracy.

[0083] Step S2, intelligent analysis, the construction intelligent safety supervision center determines whether to enter the danger zone based on the location information and the distance between the construction personnel and the operating machinery;

[0084] In step S2, the intelligent analysis method is:

[0085] The construction intelligent safety supervision center analyzes the construction site based on the location information in step S1 through the deep learning model CNN and the dynamic risk assessment model MDP;

[0086] Analyze the relative position and motion status between construction personnel and operating machinery based on position information;

[0087] In step S2, the intelligent analysis method includes:

[0088] From step S1, we get the position information P = (x, y, z) and timestamp t. For beacon i, we transform its state vector s i Defined as: Among them, s i (t) represents the state vector of beacon i at time t, x i (t),y i (t),z i (t) represents the coordinates of beacon i at time t, represents the velocity component of beacon i at time t;

[0089] Construct the time series input. For the time window T, the time series input tensor S of beacon i i Expressed as: S i =[s i (tT),s i (t-T+1),…,s i (t)], where S i represents the time series input tensor of beacon i, T represents the time window size;

[0090] Use a one-dimensional convolutional neural network to process time series data and extract feature representation. For the input sequence S i , the convolution operation is defined as:

[0091] in, represents the feature representation of beacon i after the lth layer of convolution, represents the weight matrix of the convolution kernel of the lth layer, represents the feature representation of beacon i at layer l-1, b (l) represents the convolution bias of the lth layer, σ(·) represents the nonlinear activation function ReLU, K represents the convolution kernel length, and k represents the index variable in the convolution operation;

[0092] Perform pooling operation: in, It represents the feature representation of beacon i after pooling at layer l. After multiple layers of convolution and pooling operations, the feature vector f of beacon i is finally obtained.i : where f i represents the feature vector of beacon i, and Flatten means flattening the high-dimensional feature representation into a one-dimensional vector;

[0093] The dynamic risk assessment method in step S2 is:

[0094] Define the state space S in the MDP, where each state consists of the feature vectors of all current beacons:

[0095] s(t)=[f1,f2,…,f n ], where s(t) represents the state of the system at time t, and n represents the number of beacons;

[0096] Define the action space A, where actions represent the possible actions that the system may take when a risk is detected;

[0097] The state transition probability matrix P(s′|s,a) is defined to represent the probability that the system will transfer to the next state s′ after executing action a in the current state s, which is defined as: Where P(s′|s,a) represents the transition probability from state s to s′, T(s,a) represents the change function of state s caused by action a, and Z represents the normalization constant;

[0098] Define the reward function R(s,a), which represents the immediate reward for performing action a in state s:

[0099] Among them, γ i represents the weight of beacon i, n represents the number of beacons i, g i (a) represents the expected feature vector of beacon i after executing action a, and defines the state value function V(s), which represents the maximum expected reward obtained in state s: Among them, V(s) represents the value function at state s, ∑ s′ (·) represents the sum of all possible next states;

[0100] The reinforcement learning method QLearning is used to optimize the strategy π(s) so that the action a selected in each state s maximizes the long-term reward:

[0101] Among them, π(s) represents the optimal strategy in state s;

[0102] Specifically, in the intelligent analysis stage, the construction intelligent safety supervision center uses the deep learning model CNN to process the time series data received from the UWB base station, extracts the feature representation of the beacon through multi-layer convolution and pooling operations, and effectively captures the motion status and changing trends of personnel and mechanical equipment, so that the system can accurately model the motion of each beacon. Combined with the dynamic risk assessment of the MDP model, the system can predict possible future state transfers based on the feature vectors under different states, and select the best response measures through reinforcement learning optimization strategies. The intelligent analysis mechanism greatly enhances the system's risk prediction capabilities in complex dynamic environments.

[0103] In step S2, the risk in each state is graded according to the optimized strategy π(s), and mapped to different risk levels through the value function V(s) and transmitted to the monitoring platform;

[0104] When a potential collision risk is detected in step S2, the construction intelligent safety supervision center immediately issues an alarm message through the monitoring platform;

[0105] Specifically, in actual applications, the system classifies risks into different levels based on the results of intelligent analysis, and maps different risk levels to the monitoring platform through the evaluation of the value function. When a high risk level is detected, the system can immediately issue an alarm message and use a buzzer device to issue an on-site alarm to remind relevant personnel to avoid in time or take other safety measures.

[0106] Step S3, real-time warning: once a potential collision risk is detected in step S2, an alarm message is immediately issued through the monitoring platform, and a real-time alarm is issued through a buzzer device;

[0107] Specifically, the integrated remote management function allows managers of the monitoring platform to view the location information of the construction site in real time, and to remotely call out and instruct violators, reducing the possibility of safety accidents caused by improper behavior.

[0108] Step S4, remote management, the administrator of the monitoring platform views the location information in real time, remotely calls out to the violators, and corrects the unsafe behavior in time;

[0109] Specifically, when entering the construction site where the mechanical driving collision detection system of the present invention is applied, any personnel wears a smart helmet and a buzzer, and the smart helmet has a voice function; before starting construction, set the construction personnel information, vehicle and machinery information, and set the construction radius at the same time. After starting construction work, when the personnel enter the construction area, the smart camera determines whether they are wearing a helmet bound to the beacon. This function is a prior art solution. When it is detected that the helmet bound to the beacon is not worn, an alarm is sent and the platform is monitored at the same time;

[0110] The on-site area is divided into a warning area and an alarm area. The warning area is used to detect whether there are unrelated personnel in the construction site. The alarm area is the working area. When a staff member who is not in the alarm area accidentally enters the site, an alarm will still be issued.

[0111] When a person wearing a helmet bound to a beacon is detected and enters the operation warning area, the smart camera will determine whether the person is a relevant operator. If not, a safety warning will be sent to indicate that the person has entered the area.

[0112] If the person enters the warning area, a safety alarm of unrelated person intrusion will be sent, and the driver of the operating machinery will be notified to avoid it. The monitoring platform will issue an alarm and record the entry of this person in the system.

[0113] The reason why UWB is used in the present invention is that: although GPS and RFID technologies can also be used for positioning, considering the complexity of the construction environment and the requirements for positioning accuracy, UWB technology is more suitable. Other alternatives may include lidar or visual recognition technology, but these technologies are susceptible to interference in complex environments and are relatively expensive. In comparison, UWB technology has advantages in positioning accuracy, stability and cost.

[0114] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A mechanical driving collision detection system based on UWB positioning, characterized in that: include: Positioning beacons are attached to workers including construction personnel and drivers of construction vehicles, and to operating machinery including construction machinery and equipment and construction vehicles; The UWB base station positioning system is deployed in the construction area and is responsible for receiving signals from positioning beacons and calculating their location information; The construction intelligent safety supervision center monitors and manages the construction scope, positioning beacons, and the operating scope of workers and operating machinery, and has intelligent analysis functions; Smart camera to monitor helmet wearing conditions; The monitoring platform displays the real-time location information of the positioning beacon, receives and displays alarm information, and is equipped with a buzzer device to support remote shouting function; The detection steps of the mechanical driving collision detection system based on UWB positioning include: Step S1, positioning and data transmission, the UWB base station positioning system receives the signal of the positioning beacon, calculates and transmits the location information to the construction intelligent safety supervision center; Step S2, intelligent analysis, the construction intelligent safety supervision center determines whether to enter the danger zone based on the location information and the distance between the construction personnel and the operating machinery; Step S3, real-time warning: once a potential collision risk is detected in step S2, an alarm message is immediately issued through the monitoring platform, and a real-time alarm is issued through a buzzer device; Step S4, remote management, the administrator of the monitoring platform checks the location information in real time and remotely calls out to the violators; In step S2, the intelligent analysis method is: The construction intelligent safety supervision center analyzes the construction site based on the location information in step S1 through the deep learning model CNN and the dynamic risk assessment model MDP; Analyze the relative position and motion status between construction personnel and operating machinery based on position information; In step S2, the intelligent analysis method includes: From step S1, we get the position information P = (x, y, z0 and timestamp t. For beacon i, its state vector s i Defined as: Among them, s i (t) represents the state vector of beacon i at time t, x i (t),y i (t),z i (t) represents the coordinates of beacon i at time t, represents the velocity component of beacon i at time t; Construct the time series input. For the time window T, the time series input tensor S of beacon i i Expressed as: S i =[s i (tT),s i (t-T+1),…,s i (t)], where S i represents the time series input tensor of beacon i, T represents the time window size; Use a one-dimensional convolutional neural network to process time series data and extract feature representation. For the input sequence S i , the convolution operation is defined as: in, represents the feature representation of beacon i after the lth layer of convolution, represents the weight matrix of the convolution kernel of the lth layer, represents the feature representation of beacon i at the l-1 layer, b(l) represents the convolution bias at the lth layer, σ(·) represents the nonlinear activation function ReLU, K represents the convolution kernel length, and k represents the index variable in the convolution operation; To perform pooling operation: in, It represents the feature representation of beacon i after pooling at layer l. After multiple layers of convolution and pooling operations, the feature vector f of beacon i is finally obtained. i : where f i represents the feature vector of beacon i, and Flatten means flattening the high-dimensional feature representation into a one-dimensional vector; The dynamic risk assessment method in step S2 is: Define the state space S in the MDP, where each state consists of the feature vectors of all current beacons: s(t)=[f1,f2,…,f n ], where s(t) represents the state of the system at time t, and n represents the number of beacons; Define the action space A, where actions represent the possible actions that the system may take when a risk is detected; The state transition probability matrix P(s'|s,a) is defined as the probability that the system will transfer to the next state s' after executing action a in the current state s, which is defined as: Where P(s'|s,a) represents the transition probability from state s to s', T(s,a) represents the change function of state s caused by action a, and Z represents the normalization constant; Define the reward function R(s,a), which represents the immediate reward for performing action a in state s: Among them, γ i represents the weight of beacon i, n represents the number of beacons i, g i (a) represents the expected feature vector of beacon i after executing action a, and defines the state value function V(s), which represents the maximum expected reward obtained in state s: Among them, V(s) represents the value function at state s, ∑ s '(·) represents the sum of all possible next states; The reinforcement learning method QLearning is used to optimize the strategy π(s) so that the action a selected in each state s maximizes the long-term reward: Among them, π(s) represents the optimal strategy in state s; In step S2, the risk in each state is graded according to the optimized strategy π(s), and mapped to different risk levels through the value function V(s) and transmitted to the monitoring platform; When step S2 detects a potential collision risk, the construction intelligent safety supervision center immediately sends out an alarm message through the monitoring platform.

2. A mechanical driving collision detection method based on UWB positioning as claimed in claim 1, characterized in that: In step S1, after receiving the signal of the positioning beacon, the UWB base station positioning system determines the position of the beacon and transmits the position information to the construction intelligent safety supervision center. The positioning beacon position is determined in the following way: The positioning beacon sends a signal to each base station. Assuming that n UWB base stations are deployed in the construction area, the propagation time t of the signal from the beacon to the i-th base station is i Expressed as Among them, t i represents the signal propagation time from the beacon to the i-th base station, d i represents the distance from the beacon to the i-th base station, and c represents the speed of light; Calculate the time difference between the beacon arrival at different base stations. For base station i and base station j, the arrival time difference of the signal is Δt ij for: Among them, Δt ij represents the time difference between the beacon reaching base station i and base station j, d j represents the distance from the beacon to the jth base station, Assume that the position of the beacon is (x, y, z) and the position of base station i is (x i ,y i ,z i ), the location of base station j is (x j ,y j ,z j ), then the distance difference between the two base stations is Δd ij for: Among them, (x i ,y i ,z i ) represents the known position coordinates of the i-th base station, (x j ,y j ,z j ) represents the known location coordinates of the jth base station.

3. A mechanical driving collision detection method based on UWB positioning as claimed in claim 2, characterized in that: In step S1, the method for determining the location of the positioning beacon also includes: For Δd ij Perform Taylor expansion and ignore higher-order terms to obtain the linearized equation: The approximate linear equation obtained by expansion; The linearized equations of multiple base stations are organized into matrix form, and the least squares method is used to solve the position of the beacon: AX=B,where Among them, A represents the coefficient matrix formed by the position differences of each base station, X represents the unknown position vector of the beacon, and B represents the vector formed by the product of the time difference and the speed of light.

4. A mechanical driving collision detection method based on UWB positioning as claimed in claim 3, characterized in that: In step S1, the method for determining the position of the positioning beacon also includes: calculating X, that is, the three-dimensional coordinates (x, y, z) of the beacon by the least squares method, the position information (x, y, z) and the timestamp t together constitute a data packet, and transmitting the data packet to the construction intelligent safety supervision center through the network.

Images