A fall protection device with posture self-adaptive holding and active embrace protection

Abstract

The application provides a fall protection device with posture self-adaptive holding and active holding protection, which solves the problems of falling protection and the like, and comprises a moving chassis, a semicircular support frame is installed on the moving chassis through a lifting mechanism, a posture adjusting mechanism is connected between the lifting mechanism and the semicircular support frame, the posture adjusting mechanism is used for adjusting the pitch angle of the semicircular support frame in real time according to the inclination of the slope where the device is located, so that the semicircular support frame is kept in a vertical state, controllable opening and closing semicircular mechanical arms are hinged to the top of the two sides of the semicircular support frame, respectively, retractable support rods are installed on the rear and side of the moving chassis, a sensing system for collecting human body images and detecting the inclination of the device itself is installed on the device, and the sensing system, the semicircular mechanical arms, the support rods and the posture adjusting mechanism are electrically connected with a controller. The application has the advantages of high safety, good applicability and the like.

Description

Technical Field

[0001] This invention belongs to the technical field of fall protection devices, specifically relating to a fall protection device with posture adaptive maintenance and active hugging protection. Background Technology

[0002] With the increasing aging population and growing demand for rehabilitation and nursing care, intelligent walking aids and fall protection devices have gradually become research hotspots in the field of medical and nursing devices. Currently, most walking aids on the market use a fixed frame structure, with a few products incorporating simple mechanical support components. The support structure of existing walking aids or protective frames is usually fixed to the chassis or can only tilt as a whole with the vehicle body. When users walk on sloping surfaces such as stairs or ramps, the protective frame tilts along with the vehicle body, failing to maintain an upright posture. This causes the support direction to deviate from the user's center of gravity, weakening or even eliminating the ability to protect against backward and lateral tilting.

[0003] Furthermore, common walking aids often employ passive barriers, fixed handrails, or airbags for one-way blocking, failing to proactively respond to real-time changes in the user's posture. Especially in situations involving forward or sideways leaning, they lack mechanical structures that can actively close and wrap around the user, applying a corresponding directional reaction force, resulting in inadequate protection during sudden imbalances. Simultaneously, many walking aids with following functions use fixed-parameter PID or simple tracking control, unable to dynamically adjust their following characteristics based on changes in walking speed, distance, and environmental obstacles. At low speeds, they are prone to tracking dead zones, while at high speeds, they are prone to overshoot and sudden stops and starts, interfering with the user's normal walking, and even lacking a reasonable collision avoidance and deceleration mechanism when approaching obstacles.

[0004] To address the aforementioned issues, this application proposes a fall protection device that can actively identify multi-directional tipping trends, maintain effective vertical support in complex terrains such as slopes, implement progressive buffering when the human and machine tip together, and smoothly and safely follow, featuring posture adaptive maintenance and active embrace protection. Summary of the Invention

[0005] The purpose of this invention is to address the above-mentioned problems by providing a fall protection device that is reasonably designed, provides effective support, and features posture adaptive maintenance and active hugging protection.

[0006] Another objective of this invention is to provide a fall protection method capable of actively identifying multi-directional tilting trends, addressing the aforementioned problems.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: a fall protection device with adaptive posture maintenance and active hugging protection, comprising a mobile chassis, a semi-circular support frame mounted on the mobile chassis via a lifting mechanism, and a posture adjustment mechanism connected between the lifting mechanism and the semi-circular support frame, used to adjust the pitch angle of the semi-circular support frame in real time according to the inclination of the slope where the device is located, so that the semi-circular support frame remains vertical; semi-circular robotic arms that can be controlled to open and close are respectively hinged to the top of both sides of the semi-circular support frame; telescopic support rods are installed on the rear and sides of the mobile chassis; the device is equipped with a sensing system for acquiring human body images and detecting the tilt angle of the device itself, and the sensing system, semi-circular robotic arms, support rods, and posture adjustment mechanism are electrically connected to the controller.

[0008] This device provides omnidirectional fall protection by combining active embracing, support reaction force, and ramp verticality. The mobile chassis integrates drive wheels and lidar, enabling fixed-distance following on a flat surface. The lifting mechanism automatically adjusts the height of the semi-circular support frame based on the user's waist position. The posture adjustment mechanism dynamically adjusts the support frame's pitch angle using ramp tilt feedback, ensuring the support force direction remains parallel to the gravity line. The semi-circular robotic arms extend to the sides when not in use, not hindering normal walking; they quickly close when a forward or sideways tilt is detected, inflating airbags to form a cushioning envelope. Tail and side support rods extend to the opposite side of the tilt direction, providing support force in the opposite direction to prevent a complete fall.

[0009] In the aforementioned fall protection device with posture adaptive maintenance and active hugging protection, multiple inflatable air cushions are provided on the inner side of the semi-circular robotic arms, and the ends of a pair of semi-circular robotic arms are provided with a reinforced connection structure that can be locked together. The inflatable air cushions are located on the protruding parts inside the hugging device. When the robotic arms are closed and inflated, they can form a flexible buffer layer between the human body and the robotic arms, reducing the impact on the chest and abdomen. The end reinforcement connection structure is used for mechanical locking after the two robotic arms are closed, preventing the hug from accidentally opening and ensuring that it remains stable and closed when the weight of the human body is pressed against it.

[0010] In the aforementioned fall protection device with adaptive posture maintenance and active grip protection, the posture adjustment mechanism includes a slide at a fixed height h, an electric push rod that drives the slide to move horizontally, and a connecting rod hinged between the slide and the semi-circular support frame; the controller adjusts the device's pitch angle based on the sensor system. The electric actuator is controlled to displace the slide, thereby adjusting the tilt angle of the semi-circular support frame. Adjust to θ1 = arctan(x / h) and maintain The movement of the slide changes the support point of the connecting rod, thereby changing the tilt angle of the semi-circular support frame. The geometric relationship is as follows: the slide is fixed at a height h from the bottom of the support frame; when the slide moves forward relative to its initial position... At this time, the support frame tilts forward at an angle. ; Move backward At that time, the rearward angle In actual control, the inertial measurement unit (MPU6050) in the sensing system provides real-time device pitch angle. The controller enables position or speed closed-loop control. Real-time tracking This ensures that the semi-circular support frame remains vertical and is unaffected by the slope.

[0011] In the aforementioned fall protection device with posture adaptive holding and active embrace protection, the sensing system includes a camera, and the controller uses the YOLOv11-Pose model to extract key points of the human shoulders from the image. , and key points of both hips , The angle between the line connecting the midpoints of the shoulders and the midpoints of the hips and the vertical direction is calculated as the torso tilt angle. The vertical pixel difference between the left and right shoulders is also calculated. Based on the torso tilt angle, the vertical pixel difference, and a preset threshold, the user is determined to be leaning forward, backward, left, or right.

[0012] Specifically, through the formula: ; ; ; ; Find the midpoint of the shoulders Midpoint of both hips Define the torso tilt angle as the midpoint between the shoulders. Midpoint of both hips The angle between the connecting line and the perpendicular direction is determined by the formula: ; Calculate the body's mid-axis tilt angle ,in To prevent division by zero of constants, the formula is used again: ; Vertical pixel difference between left and right shoulders The following table reflects the degree of lateral tilt and the tendency to tip over:

[0013] in This is the attitude angle threshold. For the tilt pixel threshold, The horizontal offset baseline for the shoulder and hip is shown when standing. The average value, The threshold for judging forward and backward tilt offset.

[0014] In the aforementioned fall protection device with posture adaptive holding and active embrace protection, the controller is configured to: after a pair of semi-circular robotic arms have embraced the user, monitor the roll angle of the device through a sensing system. or pitch angle When the rate of increase in roll or pitch exceeds a preset threshold, it is determined that the device and the user are tilting together; then, a PD control algorithm is employed. or The target angle of the servo is calculated, and the servo on the corresponding side support rod is driven to slowly retract the support rod to provide a gradual buffer.

[0015] The attitude sensor MPU6050 has a built-in hardware DMP that directly outputs quaternions quat[0]∼quat[3], which are then divided by 2 to the power of 30 to obtain normalized quaternions q0,q1,q2,q3. Roll angle and pitch angle The hardware solution formula is: ; If the machine tilts forward or backward, the MPU6050 sensor installed in the middle of the vehicle calculates the vehicle's tilt angle using the following formula: ; Where 57.3 is the coefficient for radian rotation angle, i.e. When a person is already surrounded and these angles are detected to increase sharply, that is... or If the threshold is exceeded, it is considered that human-machine tilt has occurred. At this time, the PD controller is used to calculate the target rotation angle of the servo motor. , , The deviation between the current angle and the expected zero degree or the angle at the previous moment. , These are the pre-set proportional and differential coefficients. Based on these, the servo slowly retracts the corresponding support rod, gradually reducing the support force and transforming the rigid support into a soft landing, thus reducing the peak damage caused by the double impact.

[0016] In the aforementioned fall protection device with adaptive posture holding and active grip protection, a lidar is installed on the mobile chassis to measure the distance between the user and the device in real time; the controller is also configured to perform fixed-distance following: based on the deviation between the distance and the preset desired distance, a desired speed is generated by an outer-loop adaptive PID controller, wherein the proportional coefficient of the outer loop is... Integral coefficient and differential coefficients It dynamically adjusts according to the user's movement speed; the inner loop uses an incremental PID controller to track the desired speed and generate smooth motor control; and a safety scaling factor is introduced based on the obstacle distance measured by the lidar, automatically reducing or stopping the movement when approaching an obstacle.

[0017] The outer loop is responsible for calculating the theoretical speed required by the robot based on higher-level instructions and real-time errors. Adaptive PID control law: The outer loop uses a positional PID, but its parameters change with the human's movement speed. Dynamic adjustment.

[0018] ; Among them, error The parameter adaptive behavior is as follows:

[0019] in This is the benchmark parameter for the proportionality coefficient. The baseline parameter for the integral coefficient, As the base parameters for the differential coefficients, after determining the physical parameters such as the robot's load and ground friction, a set of PID parameters is tuned to ensure stable tracking when the person is stationary. This is an adaptive parameter for the scaling factor. For the adaptive parameters of the integral coefficients, These are adaptive parameters for the differential coefficients.

[0020] The inner loop is responsible for tracking the desired speed of the outer loop output with high precision and smoothness. Incremental PID control law: The inner loop uses incremental PID, directly outputting the increment of the motor control quantity. .

[0021] ; in , To control the cycle. Final output: This form naturally resists integral saturation and provides continuous and smooth output commands, fundamentally avoiding severe motor vibration and ensuring the stability and safety of execution.

[0022] Security is not a separate module, but rather a unified security scaling factor. Deeply embedded control loop. Safety factor calculation: based on the distance to the nearest obstacle. calculate: ; The critical distance at which to stop (e.g., 0.2 meters). The safe distance to begin deceleration (e.g., 0.8 meters). When ≤ hour, Forced stop. Speed ​​command synthesis: Apply the safety factor to the outer loop output: .

[0023] The baseline parameters in the above adaptive algorithm , , The initial tuning can be performed using the classic engineering method of the Ziegler-Nichols critical proportionality method to ensure that the system has a stable and reliable starting point.

[0024] First, the person is brought to a standstill, then the robot activates the pure proportional following mode (i.e., set...). , ), and disable adaptive obstacle avoidance. Adjust only from small to large. until the robot is far away A sustained, constant-amplitude oscillation occurs. Record the proportional gain value at this point; this is the critical gain. Simultaneously measure the period of the oscillation; this is the critical period. .

[0025] Calculate the baseline parameters of the PID controller using the Ziegler-Nichols formula: ; ; ; Starting with this set of parameters, fine-tuning can be performed in real-world scenarios to optimize dynamic performance.

[0026] In the aforementioned fall protection device with posture adaptive holding and active hugging protection, the controller identifies the user's waist position based on the sensor system and controls the lifting mechanism to adjust the height of the semi-circular support frame so that it is aligned with the user's waist.

[0027] A fall protection method includes the following steps: S1: Acquire human images collected by the sensor system, calculate the torso tilt angle and the vertical pixel difference between the left and right shoulders, and determine whether the user has a tendency to lean forward, backward, left or right based on preset conditions. S2: When a backward tilting trend is detected, the tail support rod is extended to provide forward support force; when a forward or side tilting trend is detected, a pair of semi-circular robotic arms are brought together to form a hug, and the corresponding side support rods are extended to provide backward support force. S3: During the movement of the device, the pitch angle of the device detected by the sensor system is acquired in real time, and the attitude adjustment mechanism is controlled to adjust the tilt angle of the semi-circular support frame so that the semi-circular support frame remains vertical.

[0028] The specific calculations for step S1 are the same as described above: key points of the shoulders and hips are extracted using YOLOv11-Pose, the midpoint coordinates and torso tilt angle θpix are calculated, and forward / backward and lateral tilt are determined using the vertical difference Δy between the left and right shoulders and a preset threshold. The motion mapping in step S2 is as follows: backward tilt, tail support rod extends, pushing forward; forward / lateral tilt, robotic arm closes, same-side support rod extends, pushing backward, forming a torque opposite to the tilting direction. The vertical posture maintenance principle in step S3 is consistent with the device section; through a closed-loop sliding platform position, θ1 = arctan(x / h) = θ2 is achieved, enabling ramp adaptation.

[0029] In one of the fall protection methods described above, step S4 is included: after the semi-circular robotic arm has embraced the user, the roll angle or pitch angle of the device is monitored in real time. When the angle suddenly increases and exceeds the threshold, it is determined that the human and machine tilt together. The servo motor of the corresponding side support rod is controlled to gradually retract the support rod according to the PD algorithm to implement gradual buffering.

[0030] One of the fall protection methods mentioned above includes a height adaptive step: the position of the human waist is identified by a sensing system and compared with the height of the tail end of the semi-circular support frame. If they are inconsistent, the lifting mechanism is driven to adjust the height of the semi-circular support frame so that it is aligned with the waist. In one of the fall protection methods described above, a distance-following step is included: the real-time distance between the device and the user is obtained using a lidar, and the distance deviation is input into an outer-loop adaptive PID controller, wherein the PID parameters are dynamically adjusted according to the user's movement speed to generate a desired speed; an inner-loop incremental PID controller tracks the desired speed to drive the moving chassis; at the same time, a safety scaling factor is calculated based on the distance to obstacles to correct the desired speed. In the aforementioned fall protection method, step S3, based on the device's pitch angle θ2, moves the slide at a fixed height h via an electric push rod to generate a horizontal displacement x, thereby adjusting the tilt angle of the semi-circular support frame. =arctan(x / h), and maintain .

[0031] Compared with existing technologies, the advantages of this invention are as follows: By actively closing the semi-circular robotic arm to form an embrace, combined with the coordinated action of the retractable support rods at the tail and sides, it can simultaneously handle various fall postures such as forward, backward, left, and right tilts, achieving omnidirectional active protection. The posture adjustment mechanism dynamically adjusts the tilt angle of the semi-circular support frame using a sliding table linkage structure, based on the real-time detection device's pitch angle, ensuring it remains vertical on inclined surfaces such as stairs and ramps. This guarantees that the direction of the support force is always parallel to the gravity line, solving the problem of traditional walkers failing to provide protection when tilted on ramps. In the event of an unavoidable simultaneous fall, the controller uses a PD algorithm to control the corresponding side support rod to slowly retract, providing gradually decreasing support force. This transforms rigid support or instantaneous retraction into a soft landing process. This progressive buffering effectively reduces the peak injury caused by the combined impact of the fall and the machine pressing down, significantly improving safety. Attached Figure Description

[0032] Figure 1 This is a schematic diagram of the device structure of the present invention; Figure 2 This is a schematic diagram of another device structure of the present invention; Figure 3 This is a schematic diagram of another device structure of the present invention; Figure 4 This is a flowchart of the method of the present invention; Figure 5 This is a schematic diagram of the angle adjustment principle of the attitude adjustment mechanism of the present invention; Figure 6 This is a schematic diagram of the device of the present invention on a ramp; In the diagram, the components are: 1. mobile chassis; 2. lifting mechanism; 3. semi-circular support frame; 4. attitude adjustment mechanism; 41. slide table; 42. electric push rod; 43. connecting rod; 5. semi-circular robotic arm; 51. inflatable air cushion; 52. reinforced connection structure; and 6. support rod. Detailed Implementation

[0033] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0034] like Figure 1-6 As shown, this embodiment provides a fall protection device with posture adaptive maintenance and active hugging protection. The device includes a mobile chassis 1, a lifting mechanism 2, a semi-circular support frame 3, a posture adjustment mechanism 4, a pair of semi-circular robotic arms 5, a telescopic support rod 6, and a sensing system and controller.

[0035] The mobile chassis 1 employs a four-wheel differential drive chassis with tracks, and integrates a drive motor, battery, and lidar. The lidar is used to measure the distance between the user and the device, as well as the distance to obstacles ahead, in real time. The lifting mechanism 2, fixed above the mobile chassis 1, consists of a vertical guide rail, a stepper motor, and a conveyor belt, and is used to adjust the overall height of the device. The semi-circular support frame 3 is connected to the top of the vertical rod of the lifting mechanism 2 via a posture adjustment mechanism 4, forming the entire protective structure.

[0036] The attitude adjustment mechanism 4 is used to maintain the verticality of the semi-circular support frame 3 on a sloping road surface. Its structure includes: a slide 41 positioned at a height h above the bottom of the semi-circular support frame 3; an electric push rod 42 that drives the slide 41 to move horizontally; and a connecting rod 43 with its two ends hinged to the slide 41 and the back of the semi-circular support frame 3, respectively. The controller acquires the device's pitch angle in real time based on the inertial measurement unit MPU6050 in the sensing system. The electric push rod 42 causes the slide table 41 to have a horizontal displacement x relative to its initial position, thereby adjusting the tilt angle of the semi-circular support frame 3. Adjusted to =arctan(h / x) and perform real-time closed-loop control to maintain... .

[0037] For example, when the device is in an uphill position, the pitch angle is... When the slope is +15°, the controller drives the slide 41 forward by a distance x = h tan(15°), causing the support frame 3 to tilt forward by 15°, thus counteracting the slope's effect and keeping the support frame 3 parallel to the plumb line, i.e., vertical. Similarly, when going downhill, the slide 41 moves backward, causing... The value is negative, but it remains vertical. The movement of the slide 41 is precisely controlled by the stroke of the push rod 42, with a control cycle of 10 milliseconds.

[0038] Each of the two top ends of the semicircular support frame 3 is hinged with a semicircular robotic arm 5. The two robotic arms can be controlled to close forward or open backward around the hinge axis. The inner side of the semicircular robotic arm 5 is provided with multiple inflatable air cushions 51. The inflatable air cushions 51 quickly inflate and expand after the hugging action is completed to form a flexible buffer layer. The ends of the pair of semicircular robotic arms 5 are provided with a reinforcing connection structure 52 that can lock together. When the two arms are closed in place, the reinforcing connection structure 52 automatically locks in place to prevent the hug from accidentally popping open.

[0039] A tail support rod 6 is installed at the rear of the mobile chassis 1, and a side support rod 6 is installed on each side of the chassis. All support rods 6 are driven by an electric telescopic mechanism or a servo motor, and can extend or retract. The ends of the support rods 6 are equipped with wear-resistant and anti-slip pads to enhance friction with the ground.

[0040] The sensing system includes an RGB camera mounted on top of the semi-circular support frame 3 to capture images of the user's front or back; and an MPU6050 inertial measurement unit mounted in the middle of the mobile chassis 1 to detect the roll angle of the device body in real time. and pitch angle And a LiDAR sensor mounted on the front of the chassis. The controller is an embedded AI computing board that connects all the sensors and actuators.

[0041] Specifically, the control procedures and fall protection methods are as follows: Step S1: The camera acquires human images at a rate of 30 FPS, and the controller runs the YOLOv11-Pose model to extract the coordinates of the key points on the human shoulders. and The coordinates of the key points of the two hips are and .

[0042] Calculate the midpoint between the shoulders: , ; Midpoint of both hips: , ; Define the trunk tilt angle for: ; Where ε = 10−6 to prevent division by zero. Calculate the vertical pixel difference between the left and right shoulders. Preset threshold: Attitude angle threshold =15∘, tilt pixel threshold =30 pixels, shoulder and hip horizontal offset baseline Recorded when the user is upright The average value is determined, and in this embodiment, the user measured... =5 pixels, front and rear tilt offset threshold =8 pixels. The judgment rule is as follows:

[0043] For example, when the system detects =20∘and =20 pixels, at the same time If the value is 5 pixels, it is determined to be a forward tilting trend.

[0044] Step S2: If a backward tilting tendency is detected, the controller immediately drives the tail support rod 6 to extend rapidly, providing forward support to counteract the torque of the body tilting backward. If a forward or side tilting tendency is detected, the following actions are performed: the pair of semi-circular robotic arms 5 are controlled to quickly close from the open state, forming an embrace around the body; simultaneously, the inflatable air cushion 51 inflates instantly, covering the chest, back, and side ribs of the body; the reinforcing connection structure 52 locks in place after closing; the corresponding side support rod 6 then extends to provide backward support to prevent the body from continuing to tilt forward. If the tendency is to tilt to the left, the left support rod 6 extends; if the tendency is to tilt to the right, the right support rod 6 extends. If both forward / backward and side tilting tendencies exist simultaneously, the forward / backward tilting is responded to according to priority, and the side support is activated simultaneously.

[0045] Step S3: As the device moves or remains stationary alongside the user, the MPU6050 continuously calculates the device's pitch angle. The solution utilizes the built-in hardware DMP (Digital Motion Processor) of the MPU6050 to directly read quaternions and obtain the angle through conversion. Specifically, it reads the quaternion array quat[0]~quat[3], divides it by 2 to the power of 30 to obtain the normalized quaternions q0,q1,q2,q3. Then it calculates the pitch angle: ; Where 57.3 is the radian rotation angle coefficient 180 / π. The controller is based on... The value of the electric push rod 42 of the attitude adjustment mechanism 4 drives the slide table 41 to move, producing a displacement x = h ⋅ tan( This causes the semi-circular support frame 3 to tilt at an angle. = The semi-circular support frame 3 remains vertical at all times. The adjustment process is smooth, with an adjustment rate of 30 degrees per second, which does not affect the user's normal walking.

[0046] Step S4: When the semi-circular robotic arm 5 has embraced the user, but the user's tilting inertia is too great and causes the device to also tip over, the MPU6050 continuously monitors the device's roll angle. (Rollover) or pitch angle (Flip-over). The hardware formula for calculating the roll angle is: ; If the person being tested has already collapsed and or If the increase suddenly exceeds the threshold, it is determined that the device and the user are tipping over together. At this time, the controller triggers a progressive buffering strategy, using the PD control algorithm to calculate the target angle of the support rod 6 servo motor. :

[0047] or ; in = -0∘ (desiring the car body to be straight) = −0∘; proportional coefficient Kp=1.2, differential coefficient Kd=0.3, both are pre-tested set values. The servo motor is based on… Gradually reduce the angle and slowly retract the corresponding support rod 6, smoothly transitioning the support force from its maximum value to zero within a short 0.5-1 second to avoid secondary impact caused by rigid retraction. For example, when the device tilts to the right, the roll angle... The angle of the servo motor increases rapidly from 5° to 30° in 0.2 seconds. The PD algorithm outputs a command to change the angle of the servo motor. The right support rod 6 gradually retracts from its fully extended state. The device slowly tilts to the side along with the human body. The peak value of the resultant force on the human body can be reduced by about 40% compared to the case without buffering.

[0048] In addition, the device performs automatic height adaptation. The sensor system's camera acquires an image of the user's body and calculates the vertical coordinates of the waist midpoint using OpenPose or YOLOv11-Pose models. The controller compares these coordinates with the current height of the tail end of the semi-circular support frame 3. If the deviation exceeds 2 cm, it drives the stepper motor of the lifting mechanism 2 to rotate, which in turn moves the vertical rod up and down via a conveyor belt until the center of the semi-circular support frame 3 is level with the user's waist. After adjustment, the stepper motor locks itself to maintain the height.

[0049] The lidar ranging module outputs the user distance every 50 milliseconds based on the triangulation principle. The desired distance dtarget of the outer-loop adaptive PID controller is set to 0.8 meters. (Error) The user's real-time horizontal movement speed Obtained through distance differentiation and low-pass filtering. The desired speed is output by the outer-loop position-type PID controller. : ; The PID parameters are based on Adaptive adjustment: =0.8×(1+0.2⋅∣ |); =1+0.5⋅∣ |0.05; =0.3×(1+0.1⋅∣ |); Reference parameters , , The critical gain was obtained by tuning using the Ziegler-Nichols method. =1.33, critical period =1.2 seconds, at =0.8, =0.8 / 0.6≈1.33, needs to be recalculated: according to the formula it should be... =0.8 / (0.6)=1.33, but in practical applications, a smaller value may be used to avoid excessive values. Here, we take 0.05 as an example. The actual optimization shall prevail. This is only an example demonstration. Differential benchmark =0.125 × 0.8 × 1.2 = 0.12, fine-tuned to 0.3 based on actual conditions. Adaptive factor =0.2, =0.5, =0.1.

[0050] The inner speed loop uses incremental PID control with a control cycle of =0.02 seconds. Assume the current speed error... Then the motor control increment:

[0051] ; Pick =2.5, =0.8, =0.5. Final output PWM duty cycle .

[0052] The safety scaling factor λ is calculated from the nearest obstacle distance (obsdobs) obtained by the lidar: ; Set the stopping critical distance =0.2 meters, safe distance =0.8 meters. Then when When the distance is ≥0.8 meters, λ=1, and normal following is possible; when... When the distance is 0.5 meters, λ = (0.5 − 0.2) / (0.8 − 0.2) = 0.5, the speed is halved; when When the distance is ≤0.2 meters, λ=0, and the device stops immediately.

[0053] Final inner loop speed command This achieves both compliance and safety in following the path.

[0054] The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific embodiments or use similar methods to substitute them, without departing from the spirit of the invention or exceeding the scope defined by the appended claims.

[0055] Although this document frequently uses terms such as mobile chassis 1, lifting mechanism 2, semi-circular support frame 3, attitude adjustment mechanism 4, slide table 41, electric push rod 42, connecting rod 43, semi-circular robotic arm 5, inflatable air cushion 51, reinforcing connection structure 52, and support rod 6, the possibility of using other terms is not excluded. The use of these terms is merely for the convenience of describing and explaining the essence of this invention; interpreting them as any additional limitation would contradict the spirit of this invention.

Claims

1. A fall protection device with posture adaptive holding and active hugging protection, comprising a mobile chassis (1), wherein the mobile chassis (1) is equipped with a semi-circular support frame (3) via a lifting mechanism (2), characterized in that, The lifting mechanism (2) and the semi-circular support frame (3) are connected by an attitude adjustment mechanism (4), which is used to adjust the pitch angle of the semi-circular support frame (3) in real time according to the inclination of the slope where the device is located, so that the semi-circular support frame (3) remains vertical; the top of both sides of the semi-circular support frame (3) are respectively hinged to semi-circular mechanical arms (5) that can be opened and closed in a controlled manner; the rear and sides of the mobile chassis (1) are equipped with telescopic support rods (6); the device is equipped with a sensing system for collecting human images and detecting the tilt angle of the device itself, and the sensing system, the semi-circular mechanical arm (5), the support rod (6) and the attitude adjustment mechanism (4) are electrically connected to the controller.

2. A fall protection device with posture adaptive holding and active embrace protection according to claim 1, characterized in that, The inner side of the semi-circular robotic arm (5) is provided with multiple inflatable air cushions (51), and the ends of a pair of semi-circular robotic arms (5) are provided with a reinforced connection structure (52) that can be locked together.

3. A fall protection device with posture adaptive holding and active embrace protection according to claim 1, characterized in that, The attitude adjustment mechanism (4) includes a slide (41) at a fixed height h, an electric push rod (42) that drives the slide (41) to move horizontally, and a connecting rod (43) hinged between the slide (41) and the semi-circular support frame (3); the controller adjusts the device pitch angle obtained by the sensing system. The electric push rod (42) is controlled to cause the slide (41) to move. This will change the tilt angle of the semi-circular support frame (3). Adjusted to =arctan(x / h), and keep .

4. A fall protection device with posture adaptive holding and active embrace protection according to claim 1, characterized in that, The sensing system includes a camera. The controller uses the YOLOv11-Pose model to extract key points of the human body's shoulders and hips from the image, calculates the angle between the line connecting the midpoints of the shoulders and the midpoints of the hips and the vertical direction as the torso tilt angle, and calculates the vertical pixel difference between the left and right shoulders. Based on the torso tilt angle, the vertical pixel difference and a preset threshold, it determines whether the user is leaning forward, backward, to the left or to the right.

5. A fall protection device with posture adaptive holding and active embrace protection according to claim 1, characterized in that, The controller is configured to monitor the roll angle of the device via a sensing system after a pair of semi-circular robotic arms (5) have embraced the user. or pitch angle When the rate of increase in roll or pitch exceeds a preset threshold, it is determined that the device and the user are tilting together; then, a PD control algorithm is employed. or The target angle of the servo is calculated, and the servo of the corresponding side support rod (6) is driven to slowly retract the support rod (6) to provide a gradual buffer.

6. A fall protection device with posture adaptive holding and active embrace protection according to claim 1, characterized in that, The mobile chassis (1) is equipped with a lidar for real-time measurement of the distance between the user and the device; the controller is also configured to perform fixed-distance following: based on the deviation between the distance and the preset desired distance, a desired speed is generated by an outer-loop adaptive PID controller, wherein the proportional coefficient of the outer loop is... Integral coefficient and differential coefficients It dynamically adjusts according to the user's movement speed; the inner loop uses an incremental PID controller to track the desired speed and generate smooth motor control; and a safety scaling factor is introduced based on the obstacle distance measured by the lidar, automatically reducing or stopping the movement when approaching an obstacle.

7. A fall protection device with posture adaptive holding and active embrace protection according to claim 1, characterized in that, The controller identifies the user's waist position based on the sensor system and controls the lifting mechanism (2) to adjust the height of the semi-circular support frame (3) so that it is aligned with the user's waist.

8. A fall protection method, used in the fall protection device with posture adaptive holding and active embrace protection as described in any one of claims 1-7, characterized in that, Includes the following steps: S1: Acquire human images collected by the sensor system, calculate the torso tilt angle and the vertical pixel difference between the left and right shoulders, and determine whether the user has a tendency to lean forward, backward, left or right based on preset conditions. S2: When a backward tilting trend is detected, the tail support rod (6) is extended to provide forward support force; When it is determined that there is a tendency to lean forward or to lean to the side, control a pair of semi-circular robotic arms (5) to close together to form a hug, and extend the side support rods (6) on the corresponding side to provide rearward support force; S3: During the movement of the device, the device pitch angle detected by the sensor system is obtained in real time, and the attitude adjustment mechanism (4) is controlled to adjust the tilt angle of the semi-circular support frame (3) so that the semi-circular support frame (3) remains vertical.

9. A fall protection method according to claim 8, characterized in that, Including step S4: After the semi-circular robotic arm (5) has embraced the user, the roll angle or pitch angle of the device is monitored in real time. When the angle suddenly increases and exceeds the threshold, it is determined that the human and machine tilt together. The servo motor of the corresponding side support rod (6) is controlled to gradually retract the support rod (6) according to the PD algorithm to implement gradual buffering.

10. A fall protection method according to claim 8, characterized in that, The height adaptive step includes: identifying the position of the human waist through the sensing system and comparing it with the height of the tail end of the semi-circular support frame (3). If they are inconsistent, the lifting mechanism (2) is driven to adjust the height of the semi-circular support frame (3) so that it is aligned with the waist. It also includes a fixed-distance following step: using the laser radar to obtain the real-time distance between the device and the user, the distance deviation is input into the outer loop adaptive PID controller, where the PID parameters are dynamically adjusted according to the user's movement speed to generate the desired speed; the inner loop incremental PID controller tracks the desired speed to drive the mobile chassis (1); at the same time, the safety scaling factor is calculated based on the distance to the obstacle to correct the desired speed; Step S3: Based on the device's pitch angle The electric push rod (42) moves the slide (41) located at a fixed height h, generating a horizontal displacement x, which in turn changes the tilt angle of the semi-circular support frame (3). =arctan(x / h), and maintain .

Images