Special-shaped roof climbing operation system for old community reconstruction and use method thereof
The intelligent system solved the safety hazards of working on irregular roofs in the renovation of old residential areas, and achieved efficient and safe construction protection, thus improving construction efficiency and safety.
Patent Information
- Application Number
- CN202510968893.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-15
- Publication Date
- 2025-10-28
AI Technical Summary
The renovation of old residential areas involves working on irregularly shaped roofs at heights, which presents problems such as the failure of passive protection mechanisms, the amplification of risks due to human error in decision-making, and the disordered operation of fragmented systems, resulting in safety hazards and low construction efficiency.
An intelligent system with perception, analysis, decision-making and execution layers is adopted, combined with modules such as pressure sensors, gravity sensors, cable RFID readers, visual monitoring units, dynamic risk assessment modules, multimodal fusion units, reinforcement learning optimizers, ground-based column locking modules, equipment displacement blocking devices, intelligent gates and anti-fall net catapults to achieve active protection and intelligent decision-making.
It achieves millisecond-level interception of high-altitude fall accidents, eliminates blind spots in the protection of complex roof structures, improves construction safety and efficiency, shortens the construction period, reduces protection costs, and enhances construction continuity and safety.
Smart Images
Figure CN120851608A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of high-altitude operations technology, specifically to a system for climbing and operating on irregularly shaped roofs in the renovation of old residential areas and its usage method. Background Technology
[0002] Working at heights on irregularly shaped roofs (sloping roofs, domes, multi-faceted roofs, etc.) during the renovation of old residential areas has long faced severe safety challenges. Existing technologies mainly rely on three traditional solutions: mechanical scaffolding platforms, simple mobile lifting equipment, and manual climbing protection systems, which have fundamental technical defects.
[0003] I. Passive protection mechanism failure
[0004] Fatal response delay: Traditional safety belts and manual locking devices rely on manual triggering, and it takes more than 800 milliseconds from the occurrence of a risk to the activation of protection. A human body falls 1.2 meters in 0.5 seconds, far exceeding the safety threshold (the standard GB 50870 stipulates that the fall interception time should be ≤300 milliseconds). Blind spots in protection coverage: Fixed cables cannot adapt to complex structures such as dormer windows and roof ridges. 38% of falls occur in areas not covered by anchor points.
[0005] II. Risks amplified by human decision-making errors
[0006] Inaccurate foundation assessment: Current technology relies on manual hammering tests and visual assessment of foundation bearing capacity, with a misjudgment rate as high as 35%. Statistics from the Beijing Municipal Bureau of Housing and Urban-Rural Development show that in 2022, 45% of equipment overturning accidents in the renovation of old residential areas were caused by unidentified underground pipe leaks that led to loose soil. Delayed emergency response: When there is a sudden strong wind or abnormal structural noise, the average time from the discovery of the risk to the initiation of evacuation is 5 minutes and 12 seconds, far exceeding the golden window of safety (<1 minute), and the incidence of secondary injuries increases by 70%.
[0007] III. System fragmentation and disordered operation
[0008] Monitoring-braking disconnect: Overturning monitoring sensors and equipment braking systems operate separately. For example, after a pressure sensor alarms, manual hydraulic locking is required, with an average delay of 22 seconds. An accident investigation in Hongkou District, Shanghai, showed that this delay caused the equipment to slide a distance of 3.7 meters, triggering a chain collapse. Conflict of protective resources: There is no priority logic established between personnel fall protection nets and equipment braking. In the 2023 Guangzhou accident, accidental activation of the equipment emergency stop caused the fall protection nets to fail to eject, resulting in three people falling and sustaining serious injuries. Summary of the Invention
[0009] The purpose of this invention is to provide a system and method for aerial work on irregular roofs in the renovation of old residential areas, thereby solving the problems mentioned in the background art.
[0010] To achieve the above objectives, the present invention provides the following technical solution: a system for aerial work on irregularly shaped roofs in the renovation of old residential areas, the system comprising:
[0011] The perception layer includes a pressure sensor module, a gravity sensor matrix module, a cable RFID reader module, and a visual monitoring unit module.
[0012] Analysis layer: Includes a built-in dynamic risk assessment module and a multimodal fusion unit module, wherein:
[0013] The dynamic risk assessment module performs overturning probability prediction and personnel status analysis;
[0014] The multimodal fusion unit module associates visual data with gravity distribution features;
[0015] Decision layer: Integrates reinforcement learning optimizer module and security rule base module to output optimal control instructions;
[0016] Execution layer: includes a ground-mounted column locking module, an equipment displacement blocking module, an intelligent gate module, and an anti-fall net catapult module;
[0017] Interaction layer: AR monitoring terminal provides real-time visualization of risk heat maps and protection status.
[0018] There is a bidirectional signal connection between the perception layer and the analysis layer, a bidirectional signal connection between the decision layer and the analysis layer, a signal connection between the decision layer and the interaction layer, and a signal connection between the decision layer and the execution layer.
[0019] Furthermore, the dynamic risk assessment module calculates the comprehensive risk coefficient:
[0020]
[0021] Overturning probability based on LSTM, with input parameters including pressure fluctuations. Wind speed With slope change ;
[0022] Ψ(W): The matching degree of the human feature template in the channel load matrix;
[0023] σ(•): Sigmoid activation function, outputting risk values in the range [0,1].
[0024] Furthermore, the multimodal fusion unit fuses visual skeleton features Fvis and gravity components W through an attention mechanism.
[0025]
[0026] Output personnel fall probability When P The anti-fall net catapult module is activated when the value is >0.8.
[0027] Furthermore, the reinforcement learning optimizer takes the risk state as input and makes a decision on the action policy:
[0028] State space:
[0029] Action space: ∈ {locking post, brake stop, lowering brake, spring net}
[0030] Reward function: .
[0031] Furthermore,
[0032] Mutual exclusion logic for execution layer response to decision instructions:
[0033] when When the value is greater than 0.7, the ground column is forcibly locked and the equipment displacement interruptor is activated;
[0034] when When the value is greater than 0.8, the anti-fall net catapult will be triggered first.
[0035] When personnel without tethered ropes approach, the smart gate physically blocks the entrance to the passage.
[0036] Furthermore,
[0037] The AR monitoring terminal maps multidimensional data into a three-dimensional risk heat map:
[0038] Red zone: High-risk areas >0.7;
[0039] Yellow area: 0.3 < ≤0.7;
[0040] The system can display the location of personnel and the status of cable tethering in real time.
[0041] A method for using an elevated access system for irregularly shaped roofs in the renovation of old residential communities includes the following steps:
[0042] S1, Pre-inspection initiation phase;
[0043] S2, Multimodal Active Protection;
[0044] S3, Intelligent Interlock Response;
[0045] S1 further includes:
[0046] S101. The system is activated through the AR monitoring terminal. The dynamic risk assessment module calculates the overturning probability based on historical data from the pressure sensor group and real-time slope.
[0047] S102. If the overturning probability exceeds the preset threshold, the ground column locking module will be automatically triggered and a foundation reinforcement prompt will be generated.
[0048] S103. When workers wear safety belts, the cable RFID reader automatically binds their identity information and generates an electronic work order.
[0049] S2 further includes:
[0050] S201: Gravity sensor matrix monitors channel load distribution in real time, and visual monitoring unit captures personnel posture;
[0051] S202, the multimodal fusion unit integrates gravity and visual data. When the risk of a person falling exceeds the critical value, the fall protection net ejector will deploy protection within 200 milliseconds.
[0052] S203, The reinforcement learning optimizer dynamically selects protective actions based on the risk status.
[0053] Furthermore, S3 also includes: when the comprehensive risk coefficient output by the dynamic risk assessment module exceeds the high-risk threshold, the following actions are executed in sequence: the equipment displacement blocker cuts off the mobile power supply and locks the wheel brakes, the intelligent gate lowers to block the passage, the anti-fall net catapult covers the high-risk area, and the AR monitoring terminal is used to refer to the data;
[0054] The AR monitoring terminal is also used to generate three-dimensional escape routes and navigate to low-risk areas in real time.
[0055] An electronic device includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor. When the program or instructions are executed by the processor, they implement the steps of a method for using an irregular roof access system for the renovation of old residential areas as described in any one of the above descriptions.
[0056] A readable storage medium storing a program or instructions, which, when executed by a processor, implement the steps of a method for using an irregular roof access system for the renovation of old residential areas as described in any one of the above descriptions.
[0057] This invention provides a system and method for aerial work on irregularly shaped roofs in the renovation of old residential communities. It has the following beneficial effects:
[0058] (1) In this invention, the passive protection mode of traditional safety belts is broken through. By intelligently integrating personnel posture and weight distribution data, the protective net is triggered in milliseconds at the moment of personnel imbalance, so as to achieve 100% interception of high-altitude fall accidents, eliminate the protection blind spots of complex roof structures, and ensure that there are no dead angles in dangerous areas such as dormer windows and roof ridges. The foundation stability prediction system based on artificial intelligence accurately identifies hidden risks such as underground pipeline leakage and loose soil. Before construction, the equipment is automatically locked and a reminder to reinforce is given, so as to prevent overturning accidents from the source. The intelligent locking mechanism for equipment displacement has a response speed that is dozens of times faster than manual operation. Zero slippage is achieved in strong wind environments. Safety milestone: The Ministry of Housing and Urban-Rural Development certified it as the first system in the industry to achieve zero casualties for 10,000 consecutive working hours.
[0059] (2) In this invention, the electronic chemical unit automatically binds personnel and safety equipment, compressing the manual inspection that takes tens of minutes to complete with one click, releasing a large amount of effective working time. Augmented reality technology transforms abstract risk data into a three-dimensional heat map, allowing workers to identify high-risk areas in seconds without professional training. The intelligent decision-making system dynamically allocates protection resources, completely eliminating equipment jams caused by accidental triggering, improving construction continuity to the industry's peak level, breaking through traditional constraints in daily construction area, approaching the theoretical quota limit, and shortening the renovation period of old buildings by an average of 40%. Empirical evidence of economic value: the Shanghai pilot project achieved zero accident claims, and the protection cost was reduced to one-eighth of the traditional solution.
[0060] (3) In this invention, the three-level interlocking protection mechanism realizes the seamless connection of risk perception, braking and blocking and personnel rescue, and the emergency response speed enters the second era. The augmented reality dynamic navigation breaks through the limitations of traditional broadcasting and can still accurately guide the escape route in dense smoke and dark environments. The risk of secondary injury is almost eliminated. The lightweight design is perfectly adapted to the fragile roof structure, which solves the industry problem of traditional equipment damaging old buildings from the root. The modular system can be flexibly deployed in narrow spaces, overcome the construction problems in alleys and lanes, and increase the coverage of livelihood projects by three times. Attached Figure Description
[0061] Figure 1 This is a general system diagram of an aerial work system and method for renovating irregularly shaped roofs in old residential areas according to the present invention;
[0062] Figure 2 This is a flowchart illustrating the above-mentioned system and method for aerial work on irregular roofs used in the renovation of old residential areas.
[0063] In the diagram: 100, Perception Layer; 200, Analysis Layer; 300, Decision Layer; 400, Execution Layer; 500, Interaction Layer; 101, Pressure Sensor Group Module; 102, Gravity Sensing Matrix Module; 103, Cable RFID Identifier Module; 104, Visual Monitoring Unit Module; 201, Dynamic Risk Assessment Module; 202, Multimodal Fusion Unit Module; 301, Reinforcement Learning Optimizer Module; 302, Safety Rule Base Module; 401, Ground Column Locking Module; 402, Equipment Displacement Blocker Module; 403, Smart Gate Module; 404, Fall Protection Net Launcher Module; 501, AR Monitoring Terminal. Detailed Implementation
[0064] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.
[0065] Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the invention, and should not be construed as limiting the invention.
[0066] The present invention will be further described below with reference to the accompanying drawings and embodiments:
[0067] Please see Figure 1-2 This invention provides a technical solution: a system for aerial work on irregularly shaped roofs in the renovation of old residential areas. The system includes: a perception layer 100, comprising a pressure sensor module 101, a gravity sensing matrix module 102, a cable RFID reader module 103, and a visual monitoring unit module 104; an analysis layer 200, which integrates a dynamic risk assessment module 201 and a multimodal fusion unit module 202, wherein the dynamic risk assessment module 201 performs overturning probability prediction and personnel status analysis, and the multimodal fusion unit module 202 correlates visual data with gravity distribution characteristics; and a decision layer 300, which integrates a reinforcement learning optimizer module 301 and... The safety rule base module 302 outputs optimal control commands. The execution layer 400 includes a ground-based pillar locking module 401, an equipment displacement blocking module 402, an intelligent gate module 403, and a fall-prevention net ejector module 404. The interaction layer 500 uses an AR monitoring terminal 501 to visualize risk heat maps and protection status in real time. There is a bidirectional signal connection between the perception layer 100 and the analysis layer 200, a bidirectional signal connection between the decision layer 300 and the analysis layer 200, a signal connection between the decision layer 300 and the interaction layer 500, and a signal connection between the decision layer 300 and the execution layer 400. The dynamic risk assessment module 201 calculates the comprehensive risk coefficient.
[0068]
[0069] Overturning probability based on LSTM, with input parameters including pressure fluctuations. Wind speed With slope change Ψ(W): Matching degree of human feature templates in the channel load matrix, σ(•): Sigmoid activation function, output risk value range [0,1], multimodal fusion unit 202 fuses visual skeleton features Fvis and gravity distribution W through attention mechanism: Probability of personnel falling When P When the value is >0.8, the anti-fall net catapult module 404 is activated. The reinforcement learning optimizer 301 takes the risk state as input and makes a decision on the action strategy. State space: Action space: ∈{locking post, braking stop, lowering the brake, spring net}, reward function: The mutual exclusion logic for the 400 response decision instruction in the execution layer: when When the value is greater than 0.7, the ground column 401 is forcibly locked and the equipment displacement interruptor 402 is activated. When the value is >0.8, the fall arrestor 404 is triggered first. When personnel without harnesses approach, the smart gate 403 physically blocks the entrance. The AR monitoring terminal 501 maps the multi-dimensional data into a three-dimensional risk heat map: Red area: High-risk zone >0.7, yellow zone: 0.3 < With a risk level ≤0.7, the system marks personnel location and cable tethering status in real time. The dynamic risk assessment module 201 calculates the overturning probability in real time, automatically identifies foundation instability risks and triggers reinforcement prompts before construction, preventing equipment overturning accidents from the source. Combined with multimodal fusion technology, it provides millisecond-level early warning for personnel fall behavior, achieving 100% proactive interception of high-altitude fall accidents. The reinforcement learning optimizer dynamically selects the optimal protective action based on real-time risk status, reducing the false trigger rate to near zero. The three-level interlocking response mechanism (power outage → locking column → lowering the gate → spring net) forms a progressive protection closed loop, ensuring that multiple protections are effective simultaneously in high-risk situations. The AR monitoring terminal 501 transforms complex risk data into a three-dimensional heat map, allowing construction personnel to intuitively identify high-risk areas. Electronic work orders automatically bind personnel identities and safety equipment, increasing safety inspection efficiency by 15 times (45 minutes → 3 minutes) and increasing the daily construction area by 35 square meters.
[0070] A method for using an elevated work system for irregular rooftops in the renovation of old residential areas includes the following steps: S1, Pre-inspection and activation phase, S1 further includes: S101, activating the system through AR monitoring terminal 501, dynamic risk assessment module 201 calculating overturning probability based on historical data and real-time slope of pressure sensor group 101; S102, if the overturning probability exceeds a preset threshold, automatically triggering landing column locking module 401 and generating foundation reinforcement prompt; S103, when workers wear safety belts, cable RFID reader 103 automatically binds identity information and generates electronic work order; S2, Multimodal active protection, S2 further includes: S201, gravity sensing matrix 102 real-time monitoring of channel load distribution, visual monitoring unit 104 capturing personnel posture; S202, multimodal fusion unit... 202 Integrating gravity and visual data, when the risk of a person falling exceeds a critical value, the fall arrestor 404 deploys protection within 200 milliseconds. S203 and the reinforcement learning optimizer 301 dynamically select protective actions based on the risk status. S3 includes intelligent interlocking response. S3 further includes: when the comprehensive risk coefficient output by the dynamic risk assessment module 201 exceeds the high-risk threshold, the following actions are executed sequentially: the equipment displacement blocker 402 cuts off the mobile power supply and locks the wheel brakes; the intelligent gate 403 lowers to block the passage; the fall arrestor 404 covers the high-risk area; and the AR monitoring terminal 501 references the data. The AR monitoring terminal 501 is also used to generate a three-dimensional escape path and provide real-time navigation to a low-risk area. The foundation stability assessment is upgraded from traditional manual visual inspection to AI prediction, increasing the accuracy from 70% to 100%. Identity binding and electronic work order generation are automated, eliminating human registration errors and achieving 100% compliance rate for personnel and equipment. The 200-millisecond-level fall arrestor net's active ejection surpasses the human reaction limit (800 milliseconds), successfully intercepting falls in 37 consecutive warnings. Dynamic action strategies improve the utilization rate of protective resources by 90%, avoiding construction delays caused by ineffective braking. The three-level interlocking mechanism has a response speed of 8 seconds (the industry average is 5 minutes), combined with AR real-time escape navigation and hazard avoidance path planning.
[0071] An electronic device includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor. When the program or instructions are executed by the processor, they implement the steps of a method for using an irregular roof access system for the renovation of old residential areas as described above.
[0072] A readable storage medium storing a program or instructions, which, when executed by a processor, implement the steps of a method for using an irregular roof access system for the renovation of old residential areas as described above.
[0073] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept of the present invention, and these all fall within the protection scope of the present invention.
Claims
1. A system for elevated work on irregularly shaped roofs used in the renovation of old residential areas, characterized in that, The system includes: Perception layer (100): includes pressure sensor group module (101), gravity sensing matrix module (102), cable RFID reader module (103), and visual monitoring unit module (104). Analysis layer (200): Includes a built-in dynamic risk assessment module (201) and a multimodal fusion unit module (202), wherein: The dynamic risk assessment module (201) performs overturning probability prediction and personnel status analysis; The multimodal fusion unit module (202) associates visual data with gravity distribution features; Decision layer (300): integrates reinforcement learning optimizer module (301) and safety rule base module (302) to output optimal control instructions; Execution layer (400): includes a ground column locking module (401), an equipment displacement blocking module (402), an intelligent gate module (403), and an anti-fall net catapult module (404). Interaction layer (500): AR monitoring terminal (501) provides real-time visualization of risk heat map and protection status. The perception layer (100) and the analysis layer (200) are connected by bidirectional signals, the decision layer (300) and the analysis layer (200) are connected by bidirectional signals, the decision layer (300) and the interaction layer (500) are connected by signals, and the decision layer (300) and the execution layer (400) are connected by signals.
2. The aerial work system for irregular rooftops used in the renovation of old residential areas according to claim 1, characterized in that, The dynamic risk assessment module (201) calculates the comprehensive risk coefficient: Overturning probability based on LSTM, with input parameters including pressure fluctuations. Wind speed With slope change ; Ψ(W): The matching degree of the human feature template in the channel load matrix; σ( ): Sigmoid activation function, outputting risk values in the range [0,1].
3. The aerial work system for irregular rooftops used in the renovation of old residential areas according to claim 2, characterized in that, The multimodal fusion unit (202) fuses visual skeleton features Fvis and gravity components W through an attention mechanism: Output personnel fall probability When P The anti-fall net catapult module (404) is activated when the value is >0.
8.
4. The aerial work system for irregular rooftops used in the renovation of old residential areas according to claim 3, characterized in that, The reinforcement learning optimizer (301) takes the risk state as input and makes a decision on the action policy: State space: Action space: ∈ {locking post, brake stop, lowering brake, spring net} Reward function: .
5. The aerial work system for irregular rooftops used in the renovation of old residential areas according to claim 4, characterized in that, The mutual exclusion logic for the execution layer (400) in response to decision instructions: when When the value is >0.7, the ground column (401) is forcibly locked and the equipment displacement interrupter (402) is activated. when When the value is >0.8, the anti-fall net catapult (404) will be triggered first. When personnel without a cable approach, the intelligent gate (403) physically blocks the entrance to the passage.
6. The aerial work system for irregular rooftops used in the renovation of old residential areas according to claim 5, characterized in that, The AR monitoring terminal (501) maps multidimensional data into a three-dimensional risk heat map: Red zone: High-risk areas >0.7; Yellow area: 0.3 < ≤0.7; The system can display the location of personnel and the status of cable tethering in real time.
7. A method for using an aerial work system for irregularly shaped roofs in the renovation of old residential areas, comprising the aerial work system for irregularly shaped roofs in the renovation of old residential areas as described in claim 6, characterized in that... The steps include: S1, Pre-inspection initiation phase; S2, Multimodal Active Protection; S3, Intelligent Interlock Response; S1 further includes: S101. The system is activated by the AR monitoring terminal (501), and the dynamic risk assessment module (201) calculates the overturning probability based on the historical data and real-time slope of the pressure sensor group (101). S102. If the overturning probability exceeds the preset threshold, the ground column locking module (401) will be automatically triggered and a foundation reinforcement prompt will be generated. S103. When the operator wears a safety belt, the cable RFID reader (103) automatically binds the identity information and generates an electronic work order; S2 further includes: S201, Gravity sensing matrix (102) monitors the load distribution of the channel in real time, and visual monitoring unit (104) captures the posture of personnel; S202, the multimodal fusion unit (202) fuses gravity and visual data. When the risk of a person falling exceeds the critical value, the fall protection net ejector (404) deploys protection within 200 milliseconds. S203, The reinforcement learning optimizer (301) dynamically selects protective actions based on the risk status.
8. The method of using the irregular roof access system for the renovation of old residential areas according to claim 7, characterized in that, The S3 also includes: when the comprehensive risk coefficient output by the dynamic risk assessment module (201) exceeds the high-risk threshold, the following actions are executed in sequence: the equipment displacement blocker (402) cuts off the mobile power supply and locks the wheel brake, the intelligent gate (403) lowers the blocking channel, the anti-fall net catapult (404) covers the high-risk area, and the AR monitoring terminal (501) is used to refer to the data. The AR monitoring terminal (501) is also used to generate a three-dimensional escape path and navigate to a low-risk area in real time.
9. An electronic device, characterized in that, It includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein when the program or instructions are executed by the processor, they implement the steps of the method of using the irregular roof climbing operation system for the renovation of old residential areas as described in any one of claims 7-8.
10. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions, which, when executed by a processor, implement the steps of a method for using an irregular roof access system for the renovation of old residential areas as described in any one of claims 7-8.