Sensor optimization deployment method for taxi and bus mixed fleet
By optimizing the sensor deployment of taxis and buses, the problem of limited monitoring coverage and detection quality in smart urban environmental monitoring is solved, achieving wider coverage and more efficient resource utilization.
Patent Information
- Application Number
- CN202510296908.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-13
- Publication Date
- 2025-06-27
AI Technical Summary
In the environmental monitoring of smart cities, it is difficult for the existing technology to effectively use a hybrid fleet of taxis and buses to deploy sensors, resulting in limited monitoring coverage and detection quality.
By optimizing the sensor deployment of taxis and buses, discrete research areas and time periods according to different monitoring task requirements, select bus lines and vehicles to maximize coverage area, and use taxis to supplement the coverage areas outside the coverage line.
A wider monitoring coverage area and higher detection quality are achieved, significantly reducing the number of sensors required, improving resource utilization efficiency, and meeting the requirements of different monitoring tasks for time and space accuracy.
Smart Images
Figure CN120220451A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of environmental monitoring and ubiquitous sensing, and relates to a technical method for dynamic data collection and analysis in a smart city. In particular, it relates to a method for optimizing the deployment of sensors for a mixed fleet of taxis and buses. Background Art
[0002] The construction and development of smart cities have become a hot topic of concern in recent years and an important development direction in the global urbanization process. The key to realizing urban intelligence lies in building an efficient ubiquitous sensing system. By deploying sensors and monitoring devices in every corner of the city, key data such as traffic flow and environmental quality can be collected in real time. After intelligent analysis of these data, it provides a scientific decision-making basis for urban management, thereby optimizing resource allocation and improving the operation efficiency of the city.
[0003] In recent years, the rise of mobile sensor technology has greatly improved the coverage and accuracy of urban data collection. By using mobile carriers such as taxis, buses, and drones to configure sensors to perform environmental monitoring tasks, this "Drive-by Sensing (DS)" technology provides a flexible, economical, and efficient solution for urban management. Compared with traditional fixed sensor networks, mobile sensors can collect data in a wider spatio-temporal range and achieve real-time perception and analysis of urban dynamic information.
[0004] In the DS technology, taxis and buses are two commonly used mobile carriers. Taxis can provide flexible data collection due to their wide operating range and high degree of freedom, while buses are suitable for tasks with higher stability due to their fixed routes and high-frequency operation characteristics. For large-scale urban monitoring tasks, buses are widely used due to their wider and more stable coverage characteristics. However, buses cannot break through their own limitations and cover data outside the routes. Therefore, joint detection tasks by a mixed fleet of taxis and buses can significantly improve the covered spatial range and detection quality. How to deploy sensors for mixed vehicle types is an optimization problem that needs to be systematically considered.
[0005] At the same time, in the field of environmental monitoring, the requirements for time frequency and spatial accuracy of different monitoring tasks vary significantly. For example, air quality monitoring usually requires high-frequency data once an hour, while the monitoring requirements for the urban heat island effect can be relaxed to once every three hours. Therefore, for different environmental monitoring tasks, different sensor deployment schemes should be provided. This can not only improve the monitoring efficiency but also reduce the deployment cost of the sensing system, providing technical support for the sustainable development of smart cities. Summary of the Invention
[0006] The present invention aims to provide a method for optimizing the deployment of sensors for a mixed fleet of taxis and buses. The goal of the monitoring task is to conduct large-scale urban monitoring tasks. Based on different time accuracy and spatial accuracy requirements, the selection of taxis and buses can be optimized to achieve the maximum monitoring coverage area.
[0007] A method for optimizing the deployment of sensors for a mixed fleet of taxis and buses according to the present invention includes the following steps:
[0008] Step 1: According to the requirements of different monitoring tasks, the research area is discretized into m spatial grids N of the same size g , denoted as set G; the entire research time T is discretized into t time periods with a length of l t .
[0009] Step 2: Deployment of bus sensors.
[0010] Step 2.1: Selection of bus lines for sensor deployment.
[0011] In the research area, all bus lines are denoted as set The goal of sensor deployment is to select a set of bus lines L such that these lines can cover the maximum number of grids. The specific model is as follows:
[0012]
[0013] In the formula, x g is a 0-1 variable. If grid g is covered by the selected bus lines, then x g = 1; otherwise, x g = 0.
[0014] Constraint conditions:
[0015]
[0016] Among them, x l is a 0-1 variable. If bus line l is selected, then x l = 1; otherwise, x l = 0; δ gl indicates whether bus line l covers grid g. If line l covers grid g, then δ gl = 1; otherwise, δ gl = 0. Δ b represents the maximum number of sensors placed on the bus line. If none of the selected bus line sets cover grid g, then x g = 0. If one or more selected bus lines cover grid g, then x g takes a value of 0 or 1.
[0017] Step 2.2: Selection of bus vehicles for sensor deployment.
[0018] After obtaining the bus routes to be deployed, the deployment of bus vehicle sensors follows the following two principles to screen out the set V of eligible buses bus :
[0019] (1) Prioritize selecting the buses with the longest operating time.
[0020] (2) When the operating times are the same, select the buses with the largest operating time span.
[0021] The set of grids covered by the buses with sensors deployed in the x-th time period is denoted as The task score of the entire bus fleet n g,t is a 0-1 variable. If grid g is covered at time t, then n g,t is 1, otherwise it is 0.
[0022] Step 3: Sensor deployment for the taxi-bus mixed fleet.
[0023] After the sensor deployment of the buses is completed, for the areas to be detected outside the routes, they will be monitored by taxis.
[0024] The time-space grids are denoted as the set G(x), where x ranges from 1 to t; it is known that the set of grids covered by the buses in the x-th period is Given t sets of time-space grids G(x) and a set V of taxis, the goal of sensor deployment is to select Δ t taxis to be equipped with sensors to maximize the coverage of time-space grids in each monitoring period. The selected taxis are denoted as the set V s ∈V; for each v k ∈V, define the evaluation function: f(v k ), which quantifies the number of newly covered time-space grids after selecting v k , that is, the number of previously uncovered space grids that the selected taxi can cover; d k,i is a 0-1 variable. If taxi v k covers grid g i , then d k,i is 1, otherwise it is 0.
[0025] First, initialize the taxi set V′ = V, the selected taxi set V taxi = 0, the taxi fleet task score Φ tax and the set of time-space grids Then, the taxi-bus sensor joint optimization deployment algorithm selects the best candidate taxi c in each iteration. b , and terminates when any of the following conditions is met: The number of selected taxis reaches the specified value Δ t , or no more taxis are able to cover the remaining uncovered space-time grids.
[0026] In each iteration, the evaluation function value f(v k ), traverse the remaining grids g in the space-time grid set G′(x). If the car covers the grid, that is, d k,i is 1, f(v k )=f(v k )+1; then select f(v k ) is the best candidate taxi c with the largest value b ; If f(v b )≥1, which means that the taxi contributes to expanding the spatial coverage, then remove all taxis with v b Covered space-time grid, at the same time, v b Remove from V' and add to V taxi In the update Φ taxi The value is Φ taxi =Φ taxi +f(v b ).
[0027] Step 4: Get the maximum taxi-bus mixed fleet mixed monitoring task fleet for V joint =V bus +V taxi , the mixed fleet monitoring task score is Φ joint =Φ bus +Φ taxi .
[0028] The beneficial technical effects of the present invention are:
[0029] 1. Wide applicability: The proposed sensor deployment method can meet the requirements of different monitoring tasks for temporal accuracy and spatial accuracy, and provide optimized strategies and deployment solutions for diverse application scenarios.
[0030] 2. Significant effectiveness: Through bus-taxi joint detection, not only high-quality coverage is guaranteed, but also the number of required sensors is significantly reduced, thereby improving resource utilization efficiency.
[0031] 3. Innovative multi-vehicle hybrid monitoring: The combined detection of taxis and buses effectively overcomes the limitations of a single vehicle model, achieves a higher monitoring coverage area, and meets the needs of complex environmental monitoring tasks. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] Figure 1 This is a schematic flow diagram of the sensor optimal deployment method for a mixed fleet of taxis and buses according to the present invention. Specific implementation manner
[0033] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation methods.
[0034] The flow of a sensor optimal deployment method for a mixed fleet of taxis and buses according to the present invention is as Figure 1 shown, and includes the following steps:
[0035] Step 1: According to the requirements of different monitoring tasks, the research area is discretized into m spatial grids N of the same size, denoted as the set G; the entire research time T is discretized into t time periods with a length of l g , denoted as the set G; the entire research time T is discretized into t time periods with a length of l t . Different grid sizes represent different spatial accuracy requirements, and different lengths of time periods represent different spatial accuracy requirements.
[0036] Within each period t∈T, as long as the grid g∈G is covered by the fleet once, the score of the corresponding spatio-temporal grid can be obtained, and the final task score Φ is expressed as:
[0037]
[0038] where, n g,t is a 0-1 variable. If the grid g is covered at time t, then n g,t is 1, otherwise it is 0.
[0039] Step 2: Deployment of bus sensors.
[0040] The deployment of bus sensors aims to obtain an approximate solution of the set of bus vehicles that can achieve the highest task score Φ bus , and the sensor deployment on the bus line is divided into two steps:
[0041] Step 2.1: Selection of bus lines for sensors to be deployed.
[0042] For the selection of bus lines, in the research area, all bus lines are denoted as the set The goal of sensor deployment is to select a set of bus lines L such that these lines can cover the largest number of grids. The specific model is as follows:
[0043] The optimization goal is to maximize the number of covered grids:
[0044]
[0045] where, x gis a 0-1 variable. If grid g is covered by the selected bus line, then x g = 1; otherwise, x g = 0.
[0046] Constraint:
[0047]
[0048] Among them, x l is a 0-1 variable. If bus line l is selected, then x l = 1; otherwise, x l = 0; δ gl indicates whether bus line l covers grid g. If line l covers grid g, then δ gl = 1; otherwise, δ gl = 0, Δ b represents the maximum number of sensors installed on the bus line. If none of the selected bus lines cover grid g, then x g = 0. If one or more selected bus lines cover grid g, then x g takes a value of 0 or 1.
[0049] Step 2.2: Selection of buses for sensor deployment.
[0050] After obtaining the bus lines for sensor deployment, the deployment of bus sensors follows the following two principles to screen out the eligible bus set V bus :
[0051] (1) Give priority to selecting the bus with the longest operation time.
[0052] (2) When the operation times are the same, select the bus with the largest operation time span.
[0053] The set of grids covered by the bus with sensors deployed within the x-th time period is denoted as The task score of the entire bus fleet n g,t is a 0-1 variable. If grid g is covered at time t, then n g,t is 1; otherwise, it is 0.
[0054] Step 3: Sensor deployment for the taxi-bus mixed fleet.
[0055] After the sensors of the buses are deployed, the areas covered by the lines can be stably monitored. For the areas to be detected outside the lines, they are monitored by taxis. The present invention designs a heuristic algorithm to select taxis for sensor deployment.
[0056] In each time period, the research area is discretized into m grids of the same size. These time - space grids are denoted as the set G(x), where x ranges from 1 to t. The set of grids covered by buses in the x - th period is known as Given t sets of time - space grids G(x) and a set of taxis V, the goal of sensor deployment is to select Δ t taxis equipped with sensors to maximize the coverage of time - space grids in each monitoring period. The selected taxis are denoted as the set V s ∈V; for each v k ∈V, an evaluation function is defined: f(v k ), which quantifies the number of newly covered time - space grids after selecting v k , that is, the number of previously uncovered space grids that the newly added taxi can cover; d k,i is a 0 - 1 variable. If taxi v k covers grid g i , then d k,i is 1, otherwise it is 0.
[0057] First, initialize the set of taxis V′ = V, the set of selected taxis V taxi = 0, the task score Φ taxi＝0 of the taxi fleet and the set of time - space grids Then, in each iteration, the taxi - bus sensor joint optimization deployment algorithm selects the optimal candidate taxi v b , and terminates when any of the following conditions is met: the number of selected taxis reaches the specified value Δ t , or there are no more taxis that can cover the remaining uncovered time - space grids.
[0058] In each iteration, first calculate the evaluation function value f(v k ) of each taxi. Traverse the remaining grids g in the set of time - space grids G′(x). If the taxi covers the grid, that is, d k,i is 1, then f(v k ) = f(v k ) + 1; then select the optimal candidate taxi v k with the largest f(v b ) value; if f(v b ) ≥ 1, which means the taxi contributes to expanding the space coverage, then remove all the time - space grids covered by v b , and at the same time, remove v b from V′ and add it to V taxi , and update the value of Φ taxi to Φ taxi = Φ taxi + f(v b ).
[0059] Details of the taxi sensor deployment heuristic algorithm (continued):
[0060] Input parameters:
[0061] l t : The number of total time intervals.
[0062] G = {g1, g2, …, gm}: The set of spatial grids.
[0063] V = {v1, v2, …, vn}: The set of available taxis.
[0064] d k,i (x): The coverage of grid g by taxi v k at the x-th time period i Coverage situation.
[0065] Δ t : The maximum number of taxis allowed to be selected.
[0066] The set of spatio-temporal network grids covered by buses.
[0067] Output:
[0068] V taxi : The set of selected taxis.
[0069] Φ taxi : The task score of the taxi fleet.
[0070] Algorithm steps:
[0071] 1. Initialization:
[0072] Create the set of available taxis V′, initialized as V′ ← V.
[0073] Initialize the set of selected taxis V taxi , set it as
[0074] Initialize the task score of the taxi fleet Φ taxi , initialized as Φ taxi ← 0.
[0075] For each time period X = 1 to l t , initialize the time - space grid 2. Selection process:
[0076] When the number of selected taxis is less than or equal to Δ t , repeat the following steps:
[0077] (1) Calculate the coverage score:
[0078] For each taxi v k ∈V′:
[0079] Initialize the coverage score f(v k ) ← 0.
[0080] For each time period x = 1 to l t :
[0081] For each grid g ∈ G′(x):
[0082] If d k,i (x) ≥ 1, update f(v k ) ← f(v k ) + d k,i (x).
[0083] (2) Select the best taxi:
[0084] Find the taxi v with the maximum coverage score f(v k ) in V′. b ∈V′.
[0085] If f(v k ) ≤ 0, terminate the algorithm (the coverage cannot be further improved).
[0086] (3) Update the covered status of grids:
[0087] For each time period x = 1 to l t :
[0088] For each demand point g i ∈G′(x):
[0089] If d k,i (x) ≥ 1, remove g i from G′(x).
[0090] (4) Update the taxi set:
[0091] Remove the selected taxi v from V′ b (V′ ← V′ - {v b}}).
[0092] Add v b to the set of selected taxis V taxi (V taxi ← V taxi ∪ {v b}}).
[0093] (5) Update the taxi task score
[0094] The taxi fleet task score is incremented by f(v b) Minute (Φ taxi ←Φ taxi + f(v b ))。
[0095] 3. Output result:
[0096] Return the selected taxi set V s 。
[0097] Return the task score Φ of the taxi fleet taxi
[0098] Step 4: Obtain the maximized taxi-bus hybrid fleet for the hybrid monitoring task as V joint = V bus + V taxi , and the task score of the hybrid fleet monitoring task is Φ joint = Φ bus + Φ taxi 。
Claims
1. A sensor optimization deployment method for a mixed fleet of taxis and buses, characterized in that: The following steps are involved: Step 1: According to the requirements of different monitoring tasks, the study area is discretized into m spatial grids N of the same size. g , recorded as set G; the entire research time T is discretized into t pieces of length l t time period; Step 2: Bus sensor deployment; Step 2.1: Select the bus routes where sensors are to be deployed; In the study area, all bus routes are recorded as a set The deployment goal of the sensors is to select a set of bus routes L so that these routes can cover the maximum number of grids. The model is as follows: In the formula, x g is a 0-1 variable. If the grid g is covered by the selected bus line, then x g =1, otherwise, x g =0; Constraints: Among them, x l is a 0-1 variable. If bus line l is selected, then x l =1, otherwise, x l =0;δ gl Indicates whether bus route l covers grid g. If route l covers grid g, then δ gl =1, otherwise δ gl =0,Δ b represents the maximum number of sensors placed on the bus routes. If none of the selected bus routes covers the grid g, then x g = 0, if one or more selected bus routes cover the grid g, then x g The value is 0 or 1; Step 2.2: Select the public transport vehicles where sensors are to be deployed; After obtaining the bus routes to be deployed, the deployment of bus sensors follows the following two principles to screen out the qualified bus set V: bus : (1) Prioritize buses with the longest operating hours; (2) If the operating hours are the same, the bus with the longest operating time span is selected; The set of grids covered by the bus with sensors deployed in the xth time period is recorded as Mission score for the entire bus fleet n g,t is a 0-1 variable. If the grid g is covered within time t, then n g,t is 1, otherwise it is 0; Step 3: Taxi-bus mixed fleet sensor deployment; After the bus sensors are deployed, taxis will monitor the areas to be detected outside the route; The time-space grid is recorded as a set G(x), where x ranges from 1 to t. The set of grids covered by the bus in the xth period is Given t sets of spatiotemporal grids G(x) and a set of taxis V, the goal of sensor deployment is to select Δ t The taxis are equipped with sensors to maximize the coverage of the spatiotemporal grid in each monitoring cycle. The selected taxis are recorded as set V s ∈V; for each v k ∈V, define the evaluation function: f(v k ), quantization is in the selection of v k The number of newly added space-time grids covered later, that is, the number of previously uncovered space grids that can be covered by the newly added taxi; d k,i is a 0-1 variable. If the taxi v k After covering the grid g i , then d k,i is 1, otherwise it is 0; First, initialize the taxi set V′=V, and select the taxi set V taxi =0, taxi fleet task score Φ taxi＝0 and space-time grid collection Then, the taxi-bus sensor joint optimization deployment algorithm selects the best candidate taxi v in each iteration. b , and terminates when any of the following conditions is met: The number of selected taxis reaches the specified value Δ t , or there are no more taxis that can cover the remaining uncovered space-time grids; In each iteration, the evaluation function value f(v k ), traverse the remaining grids g in the space-time grid set G′(x). If the car covers the grid, that is, d k,i is 1, f(v k )=f(v k )+1; then select f(v k )The optimal candidate taxi v with the largest value b ; If f(v b )≥1, which means that the taxi contributes to expanding the spatial coverage, then remove all taxis with v b Covered space-time grid, at the same time, v b Remove from V' and add to V taxi In the update Φ taxi The value is Φ taxi =Φ taxi +f(v b ); Step 4: Get the maximum taxi-bus mixed fleet mixed monitoring task fleet for V joint =V bus +V taxi , the mixed fleet monitoring task score is Φ joint =Φ bus +Φ taxi .