A motion-sensing device and intelligent sensing system to assist blind people in getting around

Abstract

This invention discloses a motion-sensing device and intelligent sensing system to assist blind people in their travel, relating to the field of barrier-free assistive device technology. It includes a motion-sensing suit body, with a depth camera installed on the front chest and abdomen, and a wide-angle high-definition camera installed in the center of the chest. Through multimodal fusion of the depth-sensing camera and AI recognition, it achieves accurate identification of static obstacles, dynamic obstacles, and environmental features, avoiding the limitations of a single sensor. It employs a distributed motor vibration module, combined with differentiated encoding of vibration position, intensity, and frequency, to achieve accurate transmission of "direction + distance + obstacle type." Simultaneously, it incorporates voice prompts from a mobile phone speaker to adapt to different environmental noise scenarios. A real-time fusion algorithm of "multimodal sensing data - mobile phone navigation data" is constructed. When the navigation path conflicts with real-time obstacles, obstacle warning information is automatically prioritized, and navigation instructions are adjusted, reducing the user's information coordination burden.

Description

Technical Field

[0001] This invention relates to the field of accessible assistive devices, specifically to a somatosensory device and intelligent sensing system to assist blind people in their travel. Background Technology

[0002] In the existing technology, assistive devices for the blind are intelligent tools designed specifically for visually impaired people. They use technologies such as sensors, positioning systems, and AI algorithms to help them complete travel tasks safely and independently. Traditional assistive devices for the blind rely on a single sensor, which lacks accuracy in recognizing complex environments, easily leading to missed or false alarms and travel risks. Some devices use a single sound prompt, which is easily interfered with in noisy environments, making it difficult for visually impaired users to quickly distinguish the type of warning information. A few vibration feedback devices only vibrate at a single point, which cannot locate the warning direction and has low information transmission efficiency. In addition, in existing devices, mobile phone navigation and environmental perception devices are independent of each other, and navigation commands and real-time obstacle information are prone to conflict. Users need to manually coordinate the two information, increasing the operational burden. Furthermore, most devices rely on fixed buttons or voice commands, which are inconvenient to operate when holding objects or in noisy environments, and cannot be customized according to user habits.

[0003] Therefore, in view of this, the present invention proposes a somatosensory device and intelligent sensing system to assist blind people in traveling, in order to make up for and improve the shortcomings of the prior art. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a motion-sensing garment for the blind based on multimodal perception, mobile phone navigation fusion, motor vibration, and voice feedback. This garment is suitable for daily travel and environmental interaction scenarios for visually impaired individuals and can provide functions such as obstacle warning, path guidance, and environmental information prompts to assist blind people in their travel. This invention also provides an intelligent sensing system to solve the corresponding technical problems mentioned in the background section.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: a somatosensory device to assist blind people in traveling, comprising a somatosensory suit body, a depth camera installed on the front chest and abdomen of the somatosensory suit body, a wide-angle high-definition camera installed in the center of the chest of the somatosensory suit body, miniature DC vibration motors built into the front abdomen, front chest and back of the somatosensory suit body, a switch installed at the chest of the somatosensory suit body, reflective heat transfer printing provided on both shoulders and the back of the somatosensory suit body, adjustment parts installed at the connection between both shoulders and the front of the somatosensory suit body, and a controller built into the back of the somatosensory suit body.

[0006] Preferably, the body of the somatosensory garment includes a first fabric layer, a second fabric layer and a third fabric layer, with the second fabric layer disposed between the first fabric layer and the third fabric layer. The first fabric layer is Taslon, the second fabric layer is sandwich mesh fabric, and the third fabric layer is knitted fabric.

[0007] An intelligent sensing system for a motion-sensing device to assist blind people in traveling, the intelligent sensing system comprising a multimodal perception module, a navigation fusion module, and a distributed feedback module; The multimodal perception module is used to scan the three-dimensional contours of obstacles in front and around the body in real time using a depth camera installed on the front chest and abdomen of the body-sensing suit, generate obstacle point cloud data, identify environmental features in real time using a wide-angle high-definition camera installed in the center of the chest of the body-sensing suit, output obstacle type and position coordinates, and complement and correct with obstacle point cloud data to obtain obstacle information and send it to the navigation fusion module. The navigation fusion module is used to acquire obstacle information, obtain real-time path planning and traffic light status information through a third-party map API, construct a perception and navigation collaborative decision-making model, automatically prioritize obstacles based on obstacle information, and replan the path in combination with map data to obtain an obstacle-free path and send it to the distributed feedback module. The distributed feedback module is used to obtain the barrier-free path and convert the obstacle information into vibration commands. These commands are then transmitted to the blind person through the vibration motor array built into the motion-sensing suit, and the vibration intensity is calculated. Simultaneously, navigation is performed according to the barrier-free path, and concise voice prompts are output using a smartphone speaker.

[0008] Preferably, in the multimodal sensing module: The depth camera has a detection range of 0.3-5 meters, a refresh rate of 30Hz, and is installed on the front chest and abdomen of the body-sensing suit. The horizontal detection angle is 88°, and the detection range is 0.3-5 meters. The wide-angle high-definition camera has a resolution of 1080P and is installed in the center of the chest of the body-sensing suit. It has a built-in pre-trained AI model that supports obstacle classification into steps, walls, pedestrians, vehicles, tactile paving, and traffic lights, with a recognition accuracy of ≥95%.

[0009] As a preferred method, the specific process for obtaining obstacle information is as follows: S101. Real-time scanning of the 3D contours of obstacles using a depth camera to generate obstacle point cloud data, the formula of which is: ,in, Let N be the set of point clouds, and N be the number of valid points. These are the three-dimensional coordinates of points on the surface of the obstacle; S102. Using a wide-angle high-definition camera with a built-in pre-trained AI model, environmental features are identified in real time, and the obstacle type T and its location coordinates are output. : ; ; Where I is the input image and T is the classification result. The coordinates of the obstacle's center; S103. Fuse obstacle point cloud data, obstacle type and location coordinates, correct obstacle distance and type, obtain corrected obstacle information and mark it.

[0010] As a preferred method, the specific process for obtaining an accessible path is as follows: S201. Acquire obstacle point cloud data, obstacle type, and location coordinates. Connect to a smartphone via Bluetooth 5.3, integrate a third-party map API, and obtain real-time path planning (R) and traffic light status (S). ; ; Where D is the destination, O is the current location, and L is the intersection coordinates; S202. By constructing a perception and navigation priority determination algorithm as the basis of the perception and navigation collaborative decision-making model, automatic priority determination is performed based on obstacle information. When an obstacle conflicts with the navigation path, an early warning is automatically triggered and the path is replanned to obtain an obstacle-free path. The priority determination formula is as follows: ; in, If an obstacle enters the safe distance threshold, an obstacle warning will be output first. S203, Local cache of recent 10-minute perception data and navigation instructions When the Bluetooth disconnection time is ≤30 seconds, cached data is activated to maintain basic functions.

[0011] As a preferred option, in the distributed feedback module: The vibration motor array includes 23 miniature DC vibration motors, each with a diameter of 6mm and a height of 12mm. The vibration intensity is infinitely adjustable. They are installed on the front abdomen, chest, and back of the body-sensing suit, forming 180° obstacle avoidance coverage and 360° navigation guidance. The miniature DC vibration motors are controlled by PWM signals, and vibrations of different positions, intensities, and frequencies correspond to differentiated information encoding.

[0012] As a preferred method, the specific process for calculating vibration intensity is as follows: Obstacle information is acquired and converted into vibration commands, based on the obstacle type T and position coordinates. The system controls 23 miniature DC vibration motors to achieve differentiated coding, transmits vibrations to the blind person, and calculates the vibration intensity V using the following formula: ; Where k is the adjustment coefficient and d is the distance to the obstacle.

[0013] As a preferred approach, the specific process for outputting concise voice prompts is as follows: Obtaining an accessible path It navigates along an accessible route, using a smartphone speaker to output concise voice prompts. The volume L is linked to the ambient noise Q, and the formula is: ; Q uses the phone's microphone to detect volume in real time, preventing it from being too loud or too soft.

[0014] Compared with existing technologies, the beneficial effects of this invention are as follows: By using a depth-sensing camera and AI recognition to achieve multimodal fusion, it can accurately identify static obstacles, dynamic obstacles, and environmental features, avoiding the limitations of a single sensor. It adopts a distributed motor vibration module, combined with differentiated encoding of vibration position, intensity, and frequency, to achieve accurate transmission of "direction + distance + obstacle type". At the same time, it is equipped with voice assistance prompts from a mobile phone speaker to adapt to different environmental noise scenarios. It constructs a real-time fusion algorithm of "multimodal sensing data - mobile phone navigation data". When the navigation path conflicts with real-time obstacles, it automatically prioritizes the output of obstacle warning information and adjusts navigation instructions, reducing the user's information coordination burden. It designs an integrated button on clothing for linkage control with a mobile APP, supports custom vibration parameters, warning distance, and function switches, and is suitable for convenient operation when holding objects or in noisy environments. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the overall structure of a preferred embodiment of the present invention; Figure 2 This is a schematic diagram of the rear structure of the body of the somatosensory garment shown in this invention; Figure 3 This is a schematic diagram of the internal front structure of the body of the somatosensory garment shown in this invention; Figure 4 This is a schematic diagram of the internal rear structure of the body of the somatosensory garment shown in this invention; Figure 5 This is a schematic diagram of the structure at the connection point of the second fabric layer as shown in this invention; Figure 6 This is a schematic diagram of the intelligent sensing system structure shown in this invention.

[0016] The numbers on the map are: 1. Body of the haptic suit; 101. First fabric layer; 102. Second fabric layer; 103. Third fabric layer; 2. Depth camera; 3. Wide-angle high-definition camera; 4. Miniature DC vibration motor; 5. Switch; 6. Reflective heat transfer printing; 7. Adjustment components; 8. Controller. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Embodiments of the present invention: Please refer to Figures 1 to 6 As shown, a motion-sensing device for assisting blind people in traveling includes a motion-sensing suit body 1. A depth camera 2 is installed on the front chest and abdomen of the motion-sensing suit body 1. A wide-angle high-definition camera 3 is installed in the center of the chest of the motion-sensing suit body 1. Miniature DC vibration motors 4 are built into the front abdomen, front chest and back of the motion-sensing suit body 1. A switch 5 is installed at the chest of the motion-sensing suit body 1. Reflective heat transfer printing 6 is provided on both shoulders and the back of the motion-sensing suit body 1. Adjustment pieces 7 are installed at the connection between both shoulders and the front of the motion-sensing suit body 1. A controller 8 is built into the upper back of the motion-sensing suit body 1. The controller 8 has a built-in battery. The switch 5 is electrically connected to the controller 8. The depth camera 2, the wide-angle high-definition camera 3, and the miniature DC vibration motor 4 are all electrically connected to the controller 8. By operating the switch 5, the controller 8 can be turned on or off, thereby controlling the opening and closing of the depth camera 2, the wide-angle high-definition camera 3, and the miniature DC vibration motor 4. Switch 5 is located on the left chest of the body-sensing suit 1. It is a round physical button with a raised design, 15mm in diameter, which is easy for blind people to touch and identify. Long press to turn on / off; single click to quickly switch the recognition distance sensitivity of depth camera 2 and wide-angle high-definition camera 3, cycling through 1m / 1.5m / 2m / 2.5m / custom; double click to identify current environmental information; click 5 times in a row to trigger the emergency call function, which can be customized to contact emergency contacts or send out help signals to the surroundings. The body of the somatosensory garment 1 includes a first fabric layer 101, a second fabric layer 102 and a third fabric layer 103, and the second fabric layer 102 is disposed between the first fabric layer 101 and the third fabric layer 103. The first fabric layer 101 is Taslon, the second fabric layer 102 is sandwich mesh fabric, and the third fabric layer 103 is knitted fabric. Taslon is the outer layer fabric, sandwich mesh is the middle layer fabric, and knitted fabric is the inner layer fabric. The body of the haptic suit 1 is shaped like a work vest and is made of elastic and breathable fabric. Modular mounting slots are pre-set in key positions such as the chest, abdomen and back to fix the multimodal sensing module and navigation fusion module. The feedback motor (i.e., the miniature DC vibration motor 4) is fixed by hot pressing process to prevent the device from shifting during movement. Soft padding is set on the inside of the garment to improve the comfort of wearing it for a long time and the convenience of cleaning.

[0019] An intelligent sensing system for a motion-sensing device to assist blind people in traveling. The intelligent sensing system includes a multimodal perception module, a navigation fusion module, and a distributed feedback module. Each module interacts with data through a wired interface or wireless Bluetooth. The multimodal perception module is used to scan the three-dimensional contours of obstacles in front and around the body in real time through the depth camera 2 installed on the front chest and abdomen of the body-sensing suit 1, generate obstacle point cloud data, and identify environmental features in real time through the wide-angle high-definition camera 3 installed in the center of the chest of the body-sensing suit 1, output the obstacle type and position coordinates, and complement and correct with the obstacle point cloud data to obtain obstacle information and send it to the navigation fusion module. The navigation fusion module is used to obtain obstacle information, obtain real-time route planning and traffic light status information through third-party map APIs, build a perception and navigation collaborative decision-making model, automatically prioritize based on obstacle information, and replan the route in combination with map data to obtain an obstacle-free route and send it to the distributed feedback module. The distributed feedback module is used to obtain the barrier-free path and convert the obstacle information into vibration commands. These commands are transmitted to the blind person through the vibration motor array built into the body of the motion-sensing suit, and the vibration intensity is calculated. At the same time, the blind person is guided according to the barrier-free path, and simple voice prompts are output using the smartphone speaker.

[0020] In the multimodal sensing module: The depth camera 2 has a detection range of 0.3-5 meters, a refresh rate of 30Hz, and is installed on the front chest and abdomen of the body-sensing suit 1. The horizontal detection angle is 88° and the detection range is 0.3-5 meters. The wide-angle HD camera 3 has a resolution of 1080P and is installed in the center of the chest of the body-sensing suit 1. It has a built-in pre-trained AI model that supports obstacle classification into steps, walls, pedestrians, vehicles, blind paths, and traffic lights, with a recognition accuracy of ≥95%.

[0021] The specific process of obtaining obstacle information is as follows: S101. The depth camera 2 scans the 3D contour of the obstacle in real time to generate obstacle point cloud data, the formula of which is: ,in, Let N be the set of point clouds, and N be the number of valid points. These are the three-dimensional coordinates of points on the surface of the obstacle; S102. The wide-angle high-definition camera 3, with its built-in pre-trained AI model, identifies environmental features in real time and outputs the obstacle type T and its location coordinates. : ; ; Where I is the input image and T is the classification result. The coordinates of the obstacle's center; S103. Fuse obstacle point cloud data, obstacle type and location coordinates, correct obstacle distance and type, obtain corrected obstacle information and mark it.

[0022] The specific process of obtaining an accessible path is as follows: S201. Acquire obstacle point cloud data, obstacle type, and location coordinates. Connect to a smartphone via Bluetooth 5.3, integrate a third-party map API, and obtain real-time path planning (R) and traffic light status (S). ; ; Where D is the destination, O is the current location, and L is the intersection coordinates; S202. By constructing a perception and navigation priority determination algorithm as the basis of the perception and navigation collaborative decision-making model, automatic priority determination is performed based on obstacle information. When an obstacle conflicts with the navigation path, an early warning is automatically triggered and the path is replanned to obtain an obstacle-free path. The priority determination formula is as follows: ; in, If an obstacle enters the safe distance threshold, an obstacle warning will be output first. S203, Local cache of recent 10-minute perception data and navigation instructions When the Bluetooth disconnection time is ≤30 seconds, cached data is activated to maintain basic functions.

[0023] In the distributed feedback module: The vibration motor array includes 23 miniature DC vibration motors 4, each with a diameter of 6mm and a height of 12mm. The vibration intensity is infinitely adjustable. They are installed on the front abdomen, chest, and back of the body-sensing suit 1, forming 180° obstacle avoidance coverage and 360° navigation guidance. The miniature DC vibration motors 4 are controlled by PWM signals, and the vibrations at different positions, intensities, and frequencies correspond to differentiated information encoding.

[0024] The specific process for calculating vibration intensity is as follows: Obstacle information is acquired and converted into vibration commands, based on the obstacle type T and position coordinates. Control 23 micro DC vibration motors 4 to achieve differential coding, transmit the vibration to the blind person, and calculate the vibration intensity V, the formula of which is: ; Where, k is the adjustment coefficient and d is the distance to the obstacle.

[0025] The specific process of outputting a concise voice prompt is as follows: Obtain the obstacle-free path and navigate according to the obstacle-free path, use the smartphone speaker to output a concise voice prompt, and the volume L is linked with the ambient noise Q, the formula of which is: ; Where, Q is detected in real time through the mobile phone microphone, and the volume is automatically adjusted within the range of 50 - 90 decibels, avoiding excessive or too low volume, avoiding noise interference or excessive volume from affecting others, and supporting voice output through a Bluetooth headset.

[0026] Develop a dedicated APP and install it in the smartphone. This APP can broadcast navigation voices, obstacle prompts and AI environment recognition results, support users to customize vibration coding, set common destinations, customize physical button shortcuts, view historical travel trajectories and perception data statistics, and at the same time support firmware upgrade.

[0027] When the blind person wears the main body 1 of the somatosensory clothing and turns it on through the switch 5, continuously scan the environment to generate obstacle point cloud data; the depth camera 2 and the wide-angle high-definition camera 3 synchronously collect images, and identify the surrounding objects through the AI algorithm; combine the map data and obstacle information to plan an obstacle-free path; convert the orientation and distance information of the obstacles into vibration instructions for the vibration motors, and send the object recognition results and navigation prompts to the mobile phone speaker; the blind person can adjust the mode through the button and set parameters through the APP to achieve personalized use.

[0028] Through the multi-modal fusion of the depth perception camera and AI recognition, achieve accurate recognition of static obstacles, dynamic obstacles and environmental features, avoid the limitations of single sensors, adopt a distributed motor vibration module, combine the differential coding of vibration position, intensity and frequency, achieve accurate transmission of "direction + distance + obstacle type", and at the same time combine the voice auxiliary prompt of the mobile phone speaker to adapt to different environmental noise scenarios, construct a real-time fusion algorithm of "multi-modal perception data - mobile phone navigation data", when the navigation path conflicts with real-time obstacles, automatically give priority to outputting obstacle warning information and adjust the navigation instructions to reduce the user's information coordination burden, design the clothing integrated button and the mobile phone APP to be linked for control, support customizing vibration parameters, warning distance and function switches, and adapt to the convenient operation in the case of holding objects by hand or in a noisy environment.

[0029] Compared with the prior art: Most existing guide devices for the blind are distributed, while this blind human-sensing garment adopts a full-body wearable structure, integrating multimodal sensors, intelligent processing modules, and tactile / auditory feedback devices into one, realizing a full-link integrated design of "perception, processing, and feedback", which solves the problems of multiple perception blind spots and cumbersome operation caused by the distributed wearing of existing devices; Existing solutions typically only have a single function. This application achieves a deep integration of navigation, obstacle avoidance, and object recognition in terms of functionality: through multi-sensor data fusion algorithms, it can plan the optimal path in real time, identify obstacles in all scenes, and accurately identify surrounding objects. Compared with existing technologies, it can provide blind people with a more comprehensive perception of outdoor scenes and greatly improve their autonomy in travel. Existing devices mostly provide feedback through a single auditory prompt. This application adopts a multimodal feedback mechanism that combines touch and hearing: through the vibration motor array built into the haptic suit, information about obstacles at different locations and distances is transmitted to the blind person through differentiated vibrations. Combined with the object recognition results broadcast by voice, the blind person can understand environmental information more intuitively and quickly, solving the problems of existing auditory feedback being easily interfered with by environmental noise and the information transmission not being intuitive. Existing guide algorithms are mostly general models, which are not adaptable to complex outdoor scenarios for blind people. This application adopts a scenario-based intelligent algorithm model, which is specifically trained for different outdoor scenarios such as city streets, parks, and shopping malls. It can adaptively adjust sensor parameters and recognition logic, which significantly improves the recognition accuracy and navigation rationality in different scenarios compared with existing technologies.

[0030] In this embodiment: The navigation fusion module uses the RK3588 as its core AI computing motherboard, which integrates 8 TOPS AI computing power and is the core of the entire somatosensory suit. The motherboard integrates power management, control loop, sound and light module, attitude sensor module, compass module, wireless module and heat dissipation management. It is responsible for processing point cloud data transmitted from the environmental perception module, camera data and data from the attitude sensor, executing multimodal data processing algorithms, performing core calculations such as obstacle recognition, path planning and environmental recognition, and converting the information into motor vibration tactile information and voice broadcast information through hardware I2C interface control and wireless module. The environmental perception module uses an R132 depth sensor camera to collect environmental depth maps and transmits the data to the navigation fusion module via USB 3.0. The camera AI recognition unit consists of a 1280*720P pixel color camera and an RGB camera integrated with an R132 depth camera. It transmits real-time captured environmental video images to the RK3588 for AI recognition calculation via a USB 2.0 cable. The distributed feedback module uses a PCA9635 to control 23 miniature DC vibration motors to generate tactile feedback, and connects to a mobile phone via a wireless module to control the mobile phone to broadcast voice prompts to generate auditory feedback. The interaction module (switch 5) is a separate physical button that controls the power on / off of the motion-sensing suit and its interactive operations. It is connected to the AI ​​computing motherboard via an integrated wiring harness to transmit button status information. On Android / iOS phones, install software to perform user-customized settings.

[0031] Alternative solution to this embodiment: 1. Sensor combination implementation method: A thermal imaging sensor can be added to the original depth camera to enable it to recognize objects and avoid obstacles in low-visibility environments such as nighttime and fog, thus expanding its application scenarios. 2. Implementation method of feedback device: The original tactile vibration feedback can be replaced with bone conduction vibration feedback. By transmitting vibration information through the bones, the interference of the external environment on tactile feedback can be reduced, allowing blind people to clearly perceive environmental cues even in noisy environments. Based on the original auditory feedback, a directional audio speaker can be added to achieve directional sound propagation, enabling blind people to more accurately determine the location of sound cues. The number of vibration motors can be adjusted based on the original vibration motors to achieve tactile feedback precision adjustment, making it more accurate and clear for blind people to perceive the position of objects. 3. Wearable structure implementation method: The original full-body haptic suit can be redesigned into a modular split structure, divided into a top module, pants module, glove module, etc. Each module can be worn individually or in combination as needed, improving the flexibility of use. The original conventional fabric body-sensing clothing can be replaced with a stretchable elastic fabric, making it suitable for blind people of different body types, while ensuring that the sensors can still work normally when stretched, thus improving the comfort and universality of wearing it. The original conventional fabric body-sensing clothing can be replaced with flexible photovoltaic fabric or biothermal power generation fabric, so that it can continuously generate electricity in the outdoor environment. Combined with the intelligent energy storage module, the power can be dynamically distributed to increase the outdoor use time. 4. Algorithm Function Implementation: Based on the original navigation and obstacle avoidance function, a social scene recognition function can be added. The algorithm can be used to identify the actions and expressions of the surrounding people to help blind people judge social scenes. Based on the original object recognition function, a voice interaction customization function can be added, allowing blind people to customize the recognition voice for specific objects, thus enhancing the personalized user experience. The size of the interval and threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by those skilled in the art for each set of sample data; as long as it does not affect the ratio between the parameter and the quantized value.

[0032] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation. In the two embodiments provided in this application, it should be understood that the disclosed apparatus and system can be implemented in other ways; for example, the apparatus embodiments described above are merely illustrative, and the division of modules is merely a logical functional division. In actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed; furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of the apparatus or module may be electrical, mechanical or other forms. The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A motion-sensing device to assist blind people in traveling, characterized in that, The device includes a body-sensing suit (1), a depth camera (2) installed on the front chest and abdomen of the body-sensing suit (1), a wide-angle high-definition camera (3) installed in the center of the chest of the body-sensing suit (1), a miniature DC vibration motor (4) built into the front abdomen, front chest and back of the body-sensing suit (1), a switch (5) installed on the chest of the body-sensing suit (1), reflective heat transfer printing (6) set on the shoulders and back of the body-sensing suit (1), an adjustment piece (7) installed at the connection between the shoulders and the front of the body-sensing suit (1), and a controller (8) built into the back of the body-sensing suit (1).

2. The somatosensory device for assisting blind people in traveling according to claim 1, characterized in that, The body of the somatosensory garment (1) includes a first fabric layer (101), a second fabric layer (102) and a third fabric layer (103), and the second fabric layer (102) is disposed between the first fabric layer (101) and the third fabric layer (103). The first fabric layer (101) is Taslon, the second fabric layer (102) is sandwich mesh, and the third fabric layer (103) is knitted fabric.

3. An intelligent sensing system for a motion-sensing device assisting blind people in traveling, applied to the motion-sensing device for assisting blind people in traveling as described in claims 1-2, characterized in that, The intelligent sensing system includes a multimodal perception module, a navigation fusion module, and a distributed feedback module; The multimodal perception module is used to scan the three-dimensional contours of obstacles in front and around the body in real time by using a depth camera (2) installed on the front chest and abdomen of the body-sensing suit (1) to generate obstacle point cloud data. It also uses a wide-angle high-definition camera (3) installed in the center of the chest of the body-sensing suit (1) to identify environmental features in real time, output obstacle type and position coordinates, and supplement and correct with obstacle point cloud data to obtain obstacle information and send it to the navigation fusion module. The navigation fusion module is used to obtain obstacle information, obtain real-time path planning and traffic light status information through a third-party map API, construct a perception and navigation collaborative decision-making model, automatically prioritize based on obstacle information, and replan the path in combination with map data to obtain an obstacle-free path and send it to the distributed feedback module. The distributed feedback module is used to obtain the barrier-free path and convert the obstacle information into vibration commands. The vibration motor array built into the body of the haptic suit (1) is used to transmit the vibration to the blind person in a differentiated manner and calculate the vibration intensity. At the same time, it navigates according to the barrier-free path and outputs concise voice prompts using the smartphone speaker.

4. The intelligent sensing system of a motion-sensing device for assisting blind people in traveling according to claim 3, characterized in that, In the multimodal sensing module: The depth camera (2) has a detection range of 0.3-5 meters, a refresh rate of 30Hz, is installed on the front chest and abdomen of the body of the haptic suit (1), has a horizontal detection angle of 88°, and a detection range of 0.3-5 meters. The wide-angle high-definition camera (3) has a resolution of 1080P and is installed in the center of the chest of the body of the tactile suit (1). It has a built-in pre-trained AI model that supports obstacle classification as steps, walls, pedestrians, vehicles, blind paths, and traffic lights, and the recognition accuracy is ≥95%.

5. The intelligent sensing system of a motion-sensing device for assisting blind people in traveling according to claim 4, characterized in that, The specific process of obtaining obstacle information is as follows: S101. The obstacle's three-dimensional contour is scanned in real time using a depth camera (2) to generate obstacle point cloud data, the formula of which is: ,in, Let N be the set of point clouds, and N be the number of valid points. These are the three-dimensional coordinates of points on the surface of the obstacle; S102. Using a wide-angle high-definition camera (3) with a built-in pre-trained AI model, environmental features are identified in real time, and the obstacle type T and location coordinates are output. : ； ； Where I is the input image and T is the classification result. The coordinates of the obstacle's center; S103. Fuse obstacle point cloud data, obstacle type and location coordinates, correct obstacle distance and type, obtain corrected obstacle information and mark it.

6. The intelligent sensing system of a motion-sensing device for assisting blind people in traveling according to claim 5, characterized in that, The specific process of obtaining an accessible path is as follows: S201. Acquire obstacle point cloud data, obstacle type, and location coordinates. Connect to a smartphone via Bluetooth 5.3, integrate a third-party map API, and obtain real-time path planning (R) and traffic light status (S). ； ； Where D is the destination, O is the current location, and L is the intersection coordinates; S202. By constructing a perception and navigation priority determination algorithm as the basis of the perception and navigation collaborative decision-making model, automatic priority determination is performed based on obstacle information. When an obstacle conflicts with the navigation path, an early warning is automatically triggered and the path is replanned to obtain an obstacle-free path. The priority determination formula is as follows: ； in, If an obstacle enters the safe distance threshold, an obstacle warning will be output first. S203, Local cache of recent 10-minute perception data and navigation instructions When the Bluetooth disconnection time is ≤30 seconds, cached data is activated to maintain basic functions.

7. The intelligent sensing system of a motion-sensing device for assisting blind people in traveling according to claim 6, characterized in that, In the distributed feedback module: The vibration motor array includes 23 miniature DC vibration motors (4) with a diameter of 6 mm and a height of 12 mm. The vibration intensity is infinitely adjustable. They are installed on the front abdomen, chest, and back of the body of the body-sensing suit (1) to form 180° obstacle avoidance coverage and 360° navigation guidance. The miniature DC vibration motors (4) are controlled by PWM signals, and the vibrations of different positions, intensities, and frequencies correspond to differentiated information encoding.

8. The intelligent sensing system of a motion-sensing device for assisting blind people in traveling according to claim 7, characterized in that, The specific process for calculating vibration intensity is as follows: Obstacle information is acquired and converted into vibration commands, based on the obstacle type T and position coordinates. The system controls 23 micro DC vibration motors (4) to achieve differentiated coding, transmits vibrations to the blind person, and calculates the vibration intensity V, which is calculated using the following formula: ； Where k is the adjustment coefficient and d is the distance to the obstacle.

9. The intelligent sensing system of a motion-sensing device for assisting blind people in traveling according to claim 8, characterized in that, The specific process for outputting concise voice prompts is as follows: Obtaining an accessible path It navigates along an accessible route, using a smartphone speaker to output concise voice prompts. The volume L is linked to the ambient noise Q, and the formula is: ； Q uses the phone's microphone to detect volume in real time, preventing it from being too loud or too soft.

Images