UWB positioning deployment optimization method based on ant colony optimization algorithm

Abstract

The invention provides a UWB positioning base station deployment optimization method based on an improved ant colony optimization algorithm. The UWB positioning base station deployment optimization method based on the improved ant colony optimization algorithm aims at achieving high-precision and high-coverage-rate UWB base station layout under the obstacle constraint condition. The method comprises the following steps: firstly, setting obstacle areas, the number of base stations and optimization parameters in a positioning environment, and generating an initial deployment scheme meeting obstacle constraints; then, a new solution is generated by adopting a pheromone archiving mechanism and a roulette strategy, evaluation is carried out in combination with a multi-target fitness function containing dispersity, uniformity and coverage rate, and a Gaussian disturbance and local search strategy is introduced to enhance the global optimization ability, so that local optimum is effectively avoided; and finally, outputting a global optimal deployment scheme after iterative convergence. Experimental results show that in a typical indoor scene, compared with traditional random layout, the method has the advantages that the positioning effective coverage rate is improved by about 15%-20%, the base station distribution uniformity is improved by about 25%, the UWB base station (Anchor) is effectively prevented from falling into an obstacle area, the UWB system positioning precision and robustness are remarkably improved, and the method is suitable for high-reliability positioning system deployment in a complex environment.

Description

Technical Field

[0001] This invention relates to the field of indoor and outdoor UWB positioning, specifically to a UWB positioning deployment optimization method based on ant colony optimization algorithm. Background Technology

[0002] With the continuous development of technologies such as the Internet of Things, intelligent manufacturing, smart healthcare, and smart logistics, higher demands are being placed on high-precision, low-latency indoor positioning systems. Compared to GPS technology commonly used for outdoor positioning, the indoor environment is complex, with problems such as multipath effects, signal blockage, and interference, making it difficult for traditional positioning technologies to meet the requirements for accuracy and stability. Therefore, researching a stable, reliable, and adaptable indoor positioning technology has become a common goal for both academia and industry.

[0003] Among various indoor positioning technologies, Ultra-Wideband (UWB) positioning has attracted widespread attention due to its superior performance advantages. UWB technology transmits data over an extremely wide frequency band by emitting extremely short pulse signals, featuring high time resolution, strong anti-interference capabilities, and low power consumption. Thanks to nanosecond-level pulse signals, UWB can achieve centimeter-level high-precision positioning, making it particularly suitable for real-time, accurate tracking of personnel, equipment, robots, and other targets in indoor environments.

[0004] Despite the high accuracy and robustness of UWB positioning systems, their positioning performance is still constrained by the deployment location of UWB anchors. Too few anchors lead to increased blind spots and errors, while too many anchors waste resources and increase system complexity. Therefore, achieving a reasonable and efficient deployment of UWB anchors while ensuring positioning accuracy is one of the key issues in current indoor positioning system research.

[0005] To address these challenges, an increasing number of studies are introducing intelligent optimization algorithms to assist in the automated deployment of UWB anchors. Among these, Ant Colony Optimization (ACO), a heuristic algorithm simulating ant foraging behavior, has become an important tool in solving optimization problems due to its distributed search, strong adaptability, and global optimization capabilities. This algorithm iteratively optimizes the solution space by simulating the behavior of ants releasing and sensing pheromones during path selection, ultimately approaching the global optimum. In the deployment optimization of UWB positioning systems, ACO can be used to search for the optimal UWB anchor deployment scheme under constraints such as coverage, signal strength, and deployment cost. Summary of the Invention

[0006] This invention proposes an optimization method for UWB positioning deployment based on ant colony optimization algorithm. By combining environmental information modeling, obstacle avoidance and multi-objective optimization methods, the traditional ant colony algorithm is improved to obtain the optimal solution for UWB base station (Anchor) location deployment, thereby achieving efficient and low-cost indoor positioning effect.

[0007] This invention provides a UWB positioning and deployment optimization method based on ant colony optimization algorithm, specifically including the following steps: S1 beacon environment initialization phase: Set up the positioning environment, determine the obstacle area, set the number and coverage of UWB base stations (Anchors) and constraints, initialize the optimization parameters of the ant colony algorithm, and randomly generate a batch of initial deployment schemes that meet the obstacle constraints and spacing constraints.

[0008] S2 beacon location deployment optimization phase: The ant colony optimization algorithm is adopted to generate new solutions through pheromone archiving, roulette wheel selection and Gaussian perturbation. The solutions are evaluated by multi-objective fitness functions such as dispersion, uniformity and coverage, and the quality of the deployment scheme is further improved through local search.

[0009] In the S3 optimal location estimation phase: if the number of consecutive iterations reaches the early stopping round count and the fitness improvement is less than the early stopping threshold, the early stopping mechanism is triggered and optimization stops; if early stopping is not triggered, iteration stops after reaching the maximum number of iterations. When the ant colony optimization process stops, the globally optimal UWB base station (Anchor) deployment scheme is obtained by combining the historical best solution and the current new solution, thereby improving dispersion and uniformity while ensuring effective area coverage, and with low obstacle coverage. Finally, the deployment location and optimization performance indicators are visualized.

[0010] Furthermore, in step S1, the specific steps are as follows: S1.1: Set up the experimental environment. Within a designated positioning area of ​​length L meters and width W meters, define at least one rectangular obstacle region (bounded by coordinates [x1, y1, x2, y2]). Set the number of UWB base stations to be deployed M, the coverage radius of a single base station R, the minimum number of base stations required to cover the target point m, and the minimum spacing w between UWB base stations. Mesh the positioning area with a resolution b. Initialize the parameters of the ant colony optimization algorithm, including the pheromone archive size K, the maximum number of iterations T, the population size P, the number of early stopping rounds G, and the early stopping threshold g. The value ranges for each parameter are as follows: The number of UWB base stations M: The value ranges from 10 to 20, and is determined according to the deployment area at a density of "1 base station per 100-200 square meters". The coverage radius R of a single base station ranges from 5 to 30 meters, and the value is determined based on the following criteria: ,in, This is the receiver sensitivity, with a value range of -90dBm to -70dBm; Receive power at the target point and , This refers to the transmit power of the UWB base station. The distance between the UWB base station and the target point can be calculated using an indoor logarithmic distance path loss model, with the following formula: ,in For UWB signals at a distance Path loss at the location; For reference distance Path loss at the location, Take 1 meter, The value range is 40~50dB, with a preferred value of 45dB; The path loss index has a value range of 2.0-3.0, with a preferred value of 2.5 (2.0-2.5 for open indoor spaces and 2.5-3.5 for indoor spaces with obstacles). The shadow fading value is set within the range of 4-8 dB, with a preferred value of 6 dB. To ensure coverage stability, a value of 6 dB is used in the calculation. (That is, derived based on the worst-case scenario of no shadow fading). Therefore, when the received power at the target point is greater than or equal to the received sensitivity, the point is within the coverage area, from which we can obtain... .

[0011] Minimum number of base stations covering the target point, m: The value ranges from 2 to 5, and is determined according to the positioning accuracy requirements (m≥3 for decimeter-level positioning and m≥2 for meter-level positioning). Minimum spacing w between UWB base stations: The value range is... (R is the coverage radius), and To ensure that base stations do not overlap and that signal coverage is not excessively redundant; Grid resolution b: The value ranges from 0.2 to 2 meters, and is determined according to the size of the deployment area and the positioning accuracy requirements. The higher the accuracy requirement, the smaller b should be. The preferred value is 1 meter. Pheromones archive size K: The value ranges from 30 to 100, which is the number of optimal solutions retained in the pheromone archive; Maximum number of iterations T: The value ranges from 30 to 100, which ensures sufficient optimization while avoiding excessive optimization time; Population size P: The value ranges from 2 to 5 times the number of UWB base stations, ensuring that the population size can cover a sufficient search space, avoid getting trapped in local optima, and control computational complexity and improve optimization efficiency. The number of early-stopping wheels, G, has a range of values. This ensures that premature termination leads to insufficient optimization, while also preventing ineffective iterations that waste computing power. Early stopping threshold g: The range of values ​​is... This ensures that iterations where "adaptability improvements are meaningless" can be accurately identified, avoiding premature termination due to false triggers; S1.2: Randomly generate K UWB deployment schemes within the positioning area, and select the schemes that simultaneously satisfy "all base stations do not fall into obstacle areas" and "the distance between any two base stations is not less than the minimum distance". Sort the obtained initial positions according to fitness, where the fitness function mainly integrates evaluation indicators such as coverage, dispersion and uniformity. Store the sorted K schemes in the pheromone archive as initial values.

[0012] In step S2, the specific steps are as follows: S2.1: In each generation, perform an "ant search" operation for each "ant" in the population. First, select a basic solution. This experiment uses a roulette wheel selection method to select a basic solution from the pheromone archive, where the selection probability is an exponentially decreasing probability. in, The pressure parameter is used to adjust the probability difference of choosing the basic solution in the roulette wheel betting, and its value ranges from 0.1 to 0.5. It refers to the first A pheromone archive is ranked. Based on the above formula, the probability of each pheromone archive being selected can be obtained. Then, a set of archives is selected as the base solution through random sampling. .

[0013] After selecting a basic solution, a new solution is obtained by simulating the ant's foraging process by adding Gaussian random perturbations to the basic solution. Among them, disturbance intensity , The perturbation parameter ranges from 0.3 to 0.7. For new coordinates exceeding the deployment area boundary, a mirrored boundary correction using "absolute value + modulus operation + bounce correction" is applied to bring them back into the area. If the new solution contains base stations falling into obstacle areas or violating minimum spacing, it is corrected using "local resampling + coordinate adjustment." If multiple corrections are ineffective, the solution reverts to the base solution. S2.2: Evaluate the generated new solutions. The main evaluation metric is the fitness function, whose formula is as follows: in, To effectively cover weights, The penalty weight for the obstacle area, For dispersion weights, The weights are uniform and satisfy the following conditions: .

[0014] To obtain the ratio of the effective area or the obstacle area, the positioning area must first be gridded, and then the relevant ratio can be calculated based on the number of area points. From this, we can obtain: It represents the effective coverage rate, which is the proportion of points in the unobstructed area that meet the minimum number of UWB base station (Anchor) coverage requirements.

[0015] This indicates the obstacle coverage rate, which is the proportion of an area covered by obstacles.

[0016] The dispersion is represented by the spatial distribution breadth of the deployment points calculated using the minimum spanning tree (MST).

[0017] This indicates the uniformity of the distribution of the number of times each grid point is covered, where For the standard deviation of coverage intensity, The mean; S2.3: Merge the generated new solutions with the previous pheromone archive, then sort them by calculating fitness, leaving K new UWB deployment schemes, thereby continuously retaining high-quality solutions and eliminating low-quality solutions during the iteration process; S2.4: Based on the set local search rate For the previous archives mentioned above Each solution incorporates a small perturbation and undergoes a local search, thereby further improving the quality of the solution and preventing it from getting trapped in local optima. Local search rate. The value range is 0.2 to 0.4, which can ensure the basic quality of the improved solution and further improve the performance of the solution through local search, thus meeting the UWB deployment requirements of "focusing on high-quality solution optimization and controlling computational costs".

[0018] Furthermore, in step S3, the specific steps are as follows: During the iteration process, the global optimal fitness is monitored in real time. If the number of consecutive iterations reaches the early stopping round number G and the fitness improvement is less than the early stopping threshold g, the early stopping mechanism is triggered and optimization stops. If early stopping is not triggered, iteration stops after reaching the maximum number of iterations T. The scheme with the highest fitness is selected from the final pheromone archive as the globally optimal UWB base station deployment scheme, and the base station coordinates corresponding to this scheme are output.

[0019] The advantages of the method of this invention are: This invention presents a UWB positioning and deployment optimization method based on the Ant Colony Optimization (ACO) algorithm. The ACO algorithm, through pheromone guidance and probabilistic sampling, provides positive feedback to solutions in the deployment space, enabling effective searching in complex deployment spaces. It can escape local optima and discover UWB layout schemes with high coverage, good uniformity, and high dispersion. This invention achieves multi-objective deployment requirements by using a fitness function and setting weights for various indicators. Obstacle detection functionality is also added in the experiments to improve deployment efficiency. Attached Figure Description

[0020] Figure 1 This is a flowchart of the overall implementation of the present invention (ant colony optimization algorithm).

[0021] Figure 2 This is a simulation diagram of the deployment of UWB base stations (Anchors) according to the present invention. The blue dots represent the deployment points of UWB base stations (Anchors), the red dashed circles represent the coverage area corresponding to a single UWB base station (Anchor), the light green area represents the effective coverage area (i.e., the area covered by at least three UWB base stations (Anchors), the light red area represents the obstacle coverage area, the green solid line depicts the boundary line of the effective coverage area, and the red solid line represents the coverage boundary of the obstacle area.

[0022] Figure 3 It is a graph showing the changes in each evaluation index in the fitness function during the iteration process. Detailed Implementation

[0023] This invention relates to an optimization method for UWB positioning and deployment based on an ant colony optimization algorithm, the specific concept of which is as follows: S1: Setting up the experimental environment. This experiment will use a long... Meters, width A rectangular room measuring meters was used as the test area, and two rectangular obstacle areas were set up inside the room. The coordinates of the two points on the diagonal of the rectangles were respectively... , This experiment specifies the number of UWB base stations (Anchors). The coverage area of ​​the UWB base stations (Anchors) used in the experiment was determined based on experience. The minimum spacing between UWB base stations is meters. Meters. The experimental area was divided into... Meter-resolution gridding, specifying at least one meter Only points covered by UWB base stations (Anchors) can be considered valid points. This experiment sets the population size to [missing value]. The maximum number of iterations is Next, pheromone archive size The number of early-stopping wheels that terminated. Early stop threshold .

[0024] In the experiment, the pheromone archive size was set to [size missing]. The process involves randomly generating 50 UWB deployment schemes within the designated area, and selecting those that simultaneously satisfy the criteria of "no base stations falling into obstacle areas" and "the distance between any two base stations not being less than the minimum distance". The resulting initial locations are then sorted according to fitness, and the 50 sorted schemes are stored in a pheromone archive as initial values.

[0025] S2: In each iteration, perform an "ant search" operation for each "ant" in the population. First, select a basic solution. This experiment uses a roulette wheel selection method to select a basic solution from the pheromone archive, where the selection probability is an exponentially decreasing probability. in, To select pressure parameters, It refers to the first By ranking pheromone archives, the probability of each pheromone archive being selected can be obtained. Then, by random sampling, one set of archives is selected as the base solution. .

[0026] After selecting a basic solution, a new solution is obtained by simulating the ant's foraging process by adding Gaussian random perturbations to the basic solution. Among them, disturbance intensity , These are the perturbation parameters. For new coordinates that exceed the deployment area boundary, a mirrored boundary process using "absolute value + modulus operation + bounce correction" is used to bring them back into the area. If the new solution has cases where the base station falls into an obstacle area or violates the minimum spacing, it is corrected using "local resampling + coordinate adjustment". If multiple corrections are ineffective, the solution reverts to the base solution.

[0027] The generated new solutions are evaluated. The main evaluation metric is the fitness function, whose formula is as follows: in, To effectively cover weights, The penalty weight for the obstacle area, For dispersion weights, The weights are uniform and satisfy the following conditions: .

[0028] To obtain the ratio of the effective area or the obstacle area, the positioning area must first be gridded, and then the relevant ratio can be calculated based on the number of area points. From this, we can obtain: It represents the effective coverage rate, which is the proportion of points in the unobstructed area that meet the minimum number of UWB base station (Anchor) coverage requirements.

[0029] This indicates the obstacle coverage rate, which is the proportion of an area covered by obstacles.

[0030] The dispersion is represented by the spatial distribution breadth of the deployment points calculated using the minimum spanning tree (MST).

[0031] This indicates the uniformity of the distribution of the number of times each grid point is covered, where For the standard deviation of coverage intensity, This is the mean.

[0032] The newly generated solutions are merged with the previous pheromone archive, and then sorted by fitness calculation, leaving 50 new UWB deployment schemes. This process continuously retains high-quality solutions and eliminates low-quality solutions during iteration.

[0033] Based on the set local search rate For the previous archives mentioned above Each solution introduces a small perturbation and performs a local search, thereby further improving the quality of the solution and preventing it from getting trapped in local optima.

[0034] S3: Monitor the global optimal fitness in real time during the iteration process. If the number of consecutive iterations reaches the early stopping round number... And the improvement in adaptability means If the early stopping mechanism is triggered, the optimization will stop; if the early stopping mechanism is not triggered, the iteration will continue until the maximum number of iterations is reached. Then stop. Select the scheme with the highest fitness from the final pheromone archive as the globally optimal UWB base station deployment scheme, and output the base station coordinates corresponding to this scheme.

[0035] The final results of the experiment are shown in the following table, where Table 1 shows the final results of each evaluation index, and Table 2 shows the final deployment location of the UWB base station (Anchor) when the results were obtained.

[0036] Table 1: Final Results of Evaluation Indicators Table 2: UWB Base Station (Anchor) Deployment Locations .

Claims

1. A UWB positioning and deployment optimization method based on ant colony optimization algorithm, characterized in that, Includes the following steps: S1 Beacon Environment Initialization Phase: Within a defined positioning area of ​​length L meters and width W meters, at least one rectangular obstacle region (defined by coordinates [x1, y1, x2, y2]) is defined. The number of UWB base stations to be deployed, M, the coverage radius of a single base station, R, the minimum number of base stations required to cover the target point, m, and the minimum distance w between UWB base stations are set. The positioning area is gridded with a resolution b. The parameters of the ant colony optimization algorithm are initialized, including the pheromone archive size K, the maximum number of iterations T, the population size P, the number of early stopping rounds G, and the early stopping threshold g. K initial UWB base station deployment schemes are randomly generated, and the scheme that simultaneously satisfies "all base stations do not fall into the obstacle region" and "the distance between any two base stations is not less than the minimum distance" is selected as the initial pheromone archive. S2 beacon location deployment optimization phase: Construct a multi-objective fitness function to evaluate the quality of deployment schemes; this function is a dispersion index. Coverage indicators , With uniformity index The weighted sum of the values ​​is 1; based on the pheromone archive, a base solution is selected through a roulette wheel selection mechanism, and Gaussian perturbation is applied to the coordinates of the UWB base station (Anchor) in the base solution to generate a new solution; the generated new solution is processed to ensure that it is within the positioning area; the fitness value of the new solution is calculated, and the pheromone archive is updated according to the fitness value; a local search is performed on some excellent solutions in the pheromone archive, and their fitness is further improved by small-scale coordinate perturbation; In the S3 optimal location estimation phase: if the number of consecutive iterations reaches the early stopping round number and the fitness improvement is less than the early stopping threshold, the early stopping mechanism is triggered and optimization stops; if early stopping is not triggered, iteration stops after reaching the maximum number of iterations. The deployment scheme with the highest fitness is selected from the final pheromone archive and output as the globally optimal UWB base station deployment scheme.

2. The UWB positioning and deployment optimization method based on ant colony optimization algorithm according to claim 1, characterized in that, Step S1 involves establishing the positioning environment, identifying obstacle areas, setting the number and coverage area of ​​UWB base stations (Anchors) and constraints, initializing the optimization parameters of the ant colony algorithm, and randomly generating a batch of initial deployment schemes that satisfy obstacle and spacing constraints. In step S1, the indoor positioning area is divided into a grid according to resolution w. The specific optimization target parameters and constraints are as follows: ① Coverage constraint: The effective coverage area of ​​UWB base station (Anchor) signal shall not be less than 90%, where the effective coverage area is defined as the area covered by at least m grid points of UWB base stations; the lower the coverage of obstruction areas, the better. Beacons cannot fall within obstruction areas, but to avoid dead zones caused by obstruction areas, obstruction areas can be within the coverage range of beacon signals. ② Dispersion constraint: Dispersion is calculated using the minimum spanning tree method: First, construct the Euclidean distance matrix between all base stations, generate the minimum spanning tree and calculate the total length, then compare the total length with the maximum possible minimum spanning tree length of the base stations in the region to obtain the normalized dispersion. The normalized dispersion is not less than 0.2 to ensure that the base stations are distributed in a dispersed manner and avoid centralized deployment. ③ Uniformity constraint: Uniformity is calculated by the coverage intensity variation coefficient: count the number of base stations covered by each grid point in the non-obstruction area, calculate the ratio of the standard deviation of this number to the mean (variance coefficient), and then convert it into normalized uniformity (value range 0-1) by "1 / (1+variance coefficient)". ④ Non-overlapping constraint: It is stipulated that the Euclidean distance between any two UWB base stations shall not be less than the preset minimum spacing w.

3. The UWB positioning and deployment optimization method based on ant colony optimization algorithm according to claim 1, characterized in that, Step S2 employs the ant colony optimization algorithm, generating new solutions through pheromone archiving, roulette wheel selection, and Gaussian perturbation. The solutions are evaluated using multi-objective fitness functions such as dispersion, uniformity, and coverage, and the deployment scheme is further improved through local search. The iterative evolutionary main loop of ant colony optimization (ACO) mainly includes the following steps: S2.1 Ant Search (i.e., the new solution generation stage) mainly includes the following steps: Roulette wheel selection yields the basic solution: Set the selection pressure parameter q, calculate the selection probability of each solution in the pheromone archive, which is proportional to the negative exponent of "selection pressure parameter × solution's sorting number in the archive", and randomly select one solution as the basic solution based on probability. Gaussian perturbation: Set the perturbation intensity parameter The standard deviation of the disturbance is calculated by "disturbance strength parameter × (archive size - index of the basic solution in the archive) / archive size". Then, at each base station coordinate of the basic solution, a random variable following a normal distribution (mean of 0, standard deviation of the above calculated value) is superimposed to generate new coordinates. Boundary handling: A mirrored bounce boundary handling method is adopted. For new coordinates that exceed the deployment area boundary, "absolute value + modulo operation + bounce correction" is used to make them fall back into the area. If the new solution has base stations falling into obstacle areas or violating the minimum spacing, it is corrected by "local resampling + coordinate adjustment". If multiple corrections are ineffective, it reverts to the base solution. S2.2 Evaluation of the new solution: The fitness function is used to comprehensively evaluate the merits of each coordinate position, taking into account indicators such as the dispersion and uniformity of beacon deployment; S2.3 Pheromone Update: Merge new solutions with archived solutions, sort by fitness, retain the optimal number of pheromones, and update the pheromone archive; S2.4 Local Search: For some excellent solutions in the archive, perform small-scale perturbations on each UWB. Replace the solution with the one that has higher fitness. This can improve the local quality of the solution and prevent getting trapped in local optima.

4. The UWB positioning and deployment optimization method based on ant colony optimization algorithm according to claim 1, characterized in that, Step S3 introduces an early stopping mechanism as a termination criterion. The steps of the early stopping mechanism are as follows: S3.1 Record the global optimal fitness during the iteration process. If the difference between the current iteration's optimal fitness and the global optimal fitness is less than the early stopping threshold, accumulate 1 "no-boost count". S3.2 When the "no boost count" reaches the preset number of early stop rounds, the optimization process is terminated early to avoid invalid iterations; if the early stop is not triggered after the maximum number of iterations, it stops normally. This mechanism can accurately identify algorithm convergence nodes, terminate meaningless iterations in advance, reduce hardware computing power consumption, and shorten the cycle. When the ant colony optimization process stops, it combines the historical best solution with the current new solution to obtain the globally optimal UWB base station (Anchor) deployment scheme. This improves dispersion and uniformity while ensuring effective area coverage, and has low obstacle coverage. Finally, it outputs visualized deployment locations and optimized performance indicators.

5. An electronic device, characterized in that, The device includes a memory, a processor, and a UWB base station communication module. The memory stores a program and UWB positioning optimization process data. The program contains computer instructions for implementing the UWB positioning deployment optimization method. The process data includes basic UWB positioning parameters, ant colony algorithm parameters, and optimization process data. The processor, coupled to the memory, executes the program stored in the memory to implement all steps of the ant colony-based UWB positioning deployment optimization method according to any one of claims 1 to 4. The UWB base station communication module establishes bidirectional data interaction between the electronic device and the UWB base station.

6. A computer-readable storage medium, characterized in that, Used to store computer-readable programs or instructions, which, when executed by a processor, are capable of implementing all the steps in the ant colony-based UWB positioning and deployment optimization method described in any one of claims 1 to 4.

Images