Indoor path planning method based on fusion of improved Hereum algorithm and DWA and wearable blind guiding robot dog

Abstract

The invention discloses an indoor path planning method based on fusion of an improved He-Map algorithm and a DWA and a wearable blind guiding robot dog. A user activates a blind guiding robot dog system through a voice instruction, a stable holder is adjusted to a working state, stable operation of a sensor is ensured, and a path grid map is created by taking a user demand as a blind guiding punctuation; inputting the created path grid map into the improved He-Map algorithm, and outputting a global planning track; performing local refinement on the obtained global planning trajectory by using a DWA algorithm, generating a local collision avoidance trajectory, and obtaining an optimal trajectory; smoothing the optimal trajectory generated by the DWA algorithm by using a B spline optimization method, and generating an optimal speed and angular acceleration instruction; and when the user approaches the target point, the system provides an arrival prompt to ensure that the visually impaired person can safely arrive at the destination, and the navigation service is ended. According to the invention, visually impaired people can be smoothly guided to arrive at a destination, obstacles can be avoided in time, and safe and efficient indoor navigation experience is ensured.

Description

Technical Field

[0001] The invention relates to an indoor path planning method based on an improved Hippo algorithm fused with DWA and a wearable guide robot dog. Background Art

[0002] Currently, visually impaired people face many challenges in navigating indoor environments. Although outdoor navigation aids and methods have been developed, the unique characteristics of indoor environments, such as complex layouts, dynamically changing traffic and obstacles, and regulations that generally prohibit pets, limit the effectiveness of outdoor navigation tools in indoor environments. In addition, the navigation needs of indoor environments have not received sufficient attention and resolution, resulting in limited independent mobility of visually impaired people in indoor places such as shopping malls, office buildings, and hospitals.

[0003] In view of the problems of these prior arts, the present invention proposes an innovative wearable guide robot dog system, which provides an independent, safe and efficient indoor navigation solution for the visually impaired. By using the Hippo optimization algorithm for global path planning, the global optimality of path planning is ensured, which can not only smoothly guide the visually impaired to the destination, but also avoid obstacles in time, ensuring a safe and efficient indoor navigation experience. Summary of the invention

[0004] The present invention aims to solve the problems existing in the above-mentioned prior art and provides an indoor path planning method based on an improved Hippo algorithm fused with DWA and a wearable guide robot dog.

[0005] The technical solutions adopted in the present invention are:

[0006] An indoor path planning method based on improved Hippo algorithm and DWA, comprising the following steps:

[0007] S1: The user activates the guide robot dog system through voice commands, the stable gimbal is adjusted to the working state, the sensor is ensured to operate stably, and a path grid map is created with the user's needs as the guide target point;

[0008] S2: Improve the Hippo algorithm. The improvements are as follows:

[0009] (1) The Tent chaotic mapping iterative formula is introduced to initialize the population sequence of the Hippo algorithm;

[0010] (2) During the exploration and development phases of the Hippo algorithm, the adaptive dynamic adjustment learning factor is used to search for the optimal solution when the position is updated;

[0011] (3) Improve the adaptive adjustment based on the number of iterations when updating the position;

[0012] S3: Input the path grid map created by S1 into the improved Hippo algorithm and output the global planning trajectory;

[0013] S4: Use the DWA algorithm to locally refine the global planning trajectory, generate a local collision avoidance trajectory, and obtain the optimal trajectory;

[0014] S5: Use the B-spline optimization method to smooth the optimal trajectory generated by the DWA algorithm and generate the optimal speed and angular acceleration instructions;

[0015] S6: When the user approaches the target point, the system provides an arrival prompt to ensure that the visually impaired person can reach the destination safely and ends the navigation service.

[0016] Furthermore, the introduced Tent chaotic mapping iterative formula is defined as:

[0017]

[0018] Among them, r is the chaos parameter, and its value range is [2,4]; x n is the current iteration state, x n+1 is the state of the next iteration.

[0019] Furthermore, in the exploration phase of the Hippopotamus algorithm, when the estimated value of the dispersed Hippopotamus is reached, the Hippopotamus updates its position in the river or pond, and the specific formula is:

[0020]

[0021] in, Updated Hippopotamus positions for the exploration phase; Updated front hippo position for exploration phase;X best is the best position currently found; α is a learning factor used to control the search step size;

[0022] If the improvement of the optimal solution in the current iteration is small, the exploration ability can be enhanced by increasing the learning factor:

[0023] α new =α old ×(1+Δα)

[0024] On the contrary, if the improvement is large, the development capability can be enhanced by reducing the learning factor:

[0025] α new =α old ×(1-Δα)

[0026] Among them, Δα is the preset learning factor adjustment rate; α new is the updated learning factor; α old is the learning factor before updating;

[0027] During the exploration phase, the hippopotamus will take defensive actions to avoid being captured by predators. The update formula is:

[0028]

[0029] Among them, predator represents the position of the predator, and β is the defense factor, which is used to simulate the behavior of hippopotamus escaping from predators; Hippopotamus position after the defense phase update.

[0030] Further, in the development phase, the hippo escapes from the predator and looks for a safer location. The position update formula is:

[0031]

[0032] Among them, safe_area represents the location of the safe area, and γ is the escape factor, which is used to simulate the behavior of hippopotamus escaping from predators and looking for safe areas; Updated hippo position during the escape phase.

[0033] Furthermore, when updating the position, as the number of iterations increases, the exploration parameter is gradually reduced and the development parameter is increased so that the search near the optimal solution can be more refined in the later stage of the search.

[0034] α t =α0×e -λ

[0035] β t =β0×(1-e -λt )

[0036] Among them, α t and β t are the learning factor and balance parameter of the tth iteration, λ is the decay rate, and t is the number of iterations.

[0037] Furthermore, a cubic B-spline curve is used to smooth the path after the redundant nodes are removed. The total formula of the cubic B-spline is:

[0038]

[0039] Among them, P i is the characteristic point of the control curve; F i，k (t) is the K-order B-spline basis function; i is the characteristic point index; k is the order of the B-spline basis function. Substituting k=3 into the cubic B-spline total formula equation, the following formula is obtained to represent the cubic B-spline interpolation basis function:

[0040]

[0041] The present invention also discloses a wearable guide robot dog, comprising a robot dog body, on which a laser radar and a camera are provided, and the characteristic is that a memory and a processor are also provided on the robot dog body, the memory stores a computer program, and the processor is used to call and run the computer program stored in the memory to execute the above method, so that the robot dog body performs navigation services according to the method.

[0042] Furthermore, it also includes a quick-release wearing structure, which includes a right-angle block, a wrench, a locking tongue, a tension spring and a wearing unit. The two right-angle blocks are respectively fixed to the robot dog body and the wearing unit. A locking hole is provided on each right-angle block. The locking tongue is rotatably connected in the locking hole. The wrench is fixed to the locking tongue. Both ends of the tension spring are respectively connected to the wrench and the right-angle block where the wrench is located. The two right-angle blocks are spliced ​​with each other, and the locking tongue on each right-angle block is inserted into the locking hole on the other right-angle block.

[0043] Furthermore, each right-angle block is provided with a V-shaped hole and a cone head. After the two right-angle blocks are spliced ​​together, the V-shaped holes and the cone heads on the two right-angle blocks are plugged into each other.

[0044] Furthermore, the wearable unit includes a strap and a pad, the strap is connected to the pad, a right-angle block is passed through the strap, and an arc pad is provided on the pad.

[0045] The present invention has the following beneficial effects:

[0046] The present invention provides an independent, safe and efficient indoor navigation solution for the visually impaired. By using the improved Hippo optimization algorithm for global path planning, the global optimality of path planning is ensured. At the same time, the DWA algorithm is combined for local path planning and real-time obstacle avoidance, which improves the adaptability and safety of the system. This combination utilizes the global search capability of HOA and the local obstacle avoidance capability of DWA to achieve efficient navigation in complex environments. The present invention can not only smoothly guide the visually impaired to their destination, but also avoid obstacles in time to ensure a safe and efficient indoor navigation experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is the structural diagram of a wearable guide robot dog.

[0048] Figure 2 This is the structural diagram of a wearable guide robot dog.

[0049] Figure 3 This is the structural diagram of the wearable unit.

[0050] Figure 4 This is a structural diagram of the quick-release wearing structure.

[0051] Figure 5 Structural diagram for assembling two right-angle blocks.

[0052] Figure 6 Obstacle avoidance flow chart for navigation algorithm:

[0053] Figure 7 Flowchart for improving the global planning of the Hippo algorithm. DETAILED DESCRIPTION

[0054] The present invention will be further described below in conjunction with the accompanying drawings.

[0055] like Figure 1 As shown, the present invention is a wearable guide robot dog, including a robot dog body 3, a gimbal 11 is provided on the robot dog body, a laser radar 1 is arranged on the gimbal 11, a camera 2 is arranged on the robot dog body 3, and a memory and a processor are also provided on the robot dog body. A tracking algorithm is written in the memory (the logic of the tracking algorithm will be further described below), and the processor is used to call and run the computer program stored in the memory to execute the tracking algorithm, so that the robot dog body performs navigation services according to the method.

[0056] like Figures 2 to 5 The robot dog of the present invention adopts a wearable structure, which can be worn on the arm of the user (blind person). The present application adopts a quick-release wearable structure 4, which is very easy to take the robot dog and increases the convenience of use.

[0057] The quick-release wearing structure 4 includes a right-angle block 41, a wrench 42, a locking tongue 43, a tension spring and a wearing unit 44 (for the convenience of drawing, the tension spring in the drawings of this application is not drawn). The two right-angle blocks 41 are fixed to the robot dog body 3 and the wearing unit (44) respectively. A locking hole 410 is provided on each right-angle block 41. The locking tongue 43 is rotatably connected in the locking hole 410. The wrench 42 is fixed to the locking tongue 43. The two ends of the tension spring are respectively connected to the wrench 42 and the right-angle block 41 where the wrench is located. The two right-angle blocks 41 are spliced ​​together to form a rectangular plate. The locking tongue 43 on each right-angle block 41 is inserted into the locking hole on the other right-angle block 41 to achieve the limited fixation of the two right-angle blocks 41. When it is necessary to separate the two right-angle blocks 41, pull the wrench 42 outward (such as Figure 5 ) so that the locking tongue 43 is at the right angle block 41 of the other party.

[0058] The locking tongue 43 is semicircular, and its outer arc surface is fixedly connected to the wrench 42. Figure 5 To remove the schematic diagram of the upper cover plate of the right-angle block, when the two right-angle blocks 41 are locked with each other, the locking tongue 43 is inserted into the locking hole of the opposite right-angle block 41. At this time, the two locking tongues 43 are spliced ​​to form a circular block and placed in the two spliced ​​locking holes, thereby positioning and fixing the two right-angle blocks.

[0059] In order to ensure the positioning effect of the two right-angle blocks 41 and facilitate the rapid alignment of the two right-angle blocks, a V-shaped hole 411 and a cone head 412 are provided on each right-angle block 41. After the two right-angle blocks 41 are spliced ​​together, the V-shaped holes and the cone heads on the two right-angle blocks are plugged into each other.

[0060] The wearing unit 44 in the present invention includes a strap 441 and a pad 442, wherein the strap 441 is connected to the pad 442, wherein a right-angle block 41 is inserted into the strap 441, and the strap 441 is provided with a buckle and a Japanese buckle for adjusting the size. An arc pad 443 is provided on the pad 442. The arc pad 443 is convenient for better fitting with the arm, and the pad 442 and the arc pad 443 are both made of rubber or silicone to ensure comfort of wearing. A grip rod 444 can also be provided on the pad 442, and after wearing, the hand can grasp the grip rod 444.

[0061] The robot dog of the present application can be used without being taken off. When it is used, it can be worn directly on the arm, and the handle 444 can be held in the hand. Since the robot dog body is provided with a laser radar 1 and a camera, and the laser radar 1 is arranged on a gimbal, the entire robot dog can be used as a "handheld" tracking and navigation device.

[0062] Combination Figure 6 and Figure 7 When in use, the wearable guide robot dog performs the guide task according to the following steps:

[0063] Step 1: The user activates the guide robot dog through voice commands, adjusts the stable gimbal to the working state, and ensures that the sensor operates stably.

[0064] Initialize the hippo population, each hippo represents a possible navigation path. Formula:

[0065] Stable state = f (PTZ adjustment parameter)

[0066] The laser radar and the camera are started synchronously to collect surrounding environment data. The integration of environmental data is expressed as:

[0067] Environmental data = LiDAR + camera data

[0068] χ i =lb+r×(ub-lb)

[0069] Where: i is the position of the hippopotamus, lb and ub are the lower and upper bounds of the decision variable respectively, and r is a random number between 0 and 1.

[0070] Step 2: The user specifies the destination by voice, the GPS of the guide dog locates the starting point and the destination point, and uses the laser radar on board to build a real-time map of the surrounding environment. The map building process can be expressed as:

[0071] Map = LiDAR scan × time.

[0072] Step 3: The system uses the adaptive dynamic adjustment Hippo Optimization Algorithm (adaptive HOA), that is, the improved Hippo algorithm of the present application, to calculate the optimal navigation path according to the environment map and the parameters set by the user.

[0073] Adaptive HOA simulates the social behavior of hippos and their defense strategies against predators, dynamically adjusting the learning factor and the exploration and development balance parameters to adapt to different stages of the search process. The algorithm iteratively updates the position of each member of the hippo group and finds the global optimal path through adaptive adjustment strategies. The HOA iteration formula is:

[0074] χ i，new = x i，old +Δχ

[0075] Among them, Δχ is the position adjustment calculated based on the hippopotamus’ social behavior and predator defense strategy, which changes dynamically in each iteration according to the algorithm’s adaptive mechanism.

[0076] Step 4: Use the DWA algorithm to locally refine the global path calculated by HOA to generate a local collision avoidance trajectory. The DWA algorithm samples in the velocity space, simulates the motion trajectory, and selects the optimal trajectory based on the evaluation function to achieve local obstacle avoidance.

[0077] Step 5: Integrate the global path calculated by HOA and the local path refined by DWA algorithm to form a complete navigation path. The integrated path is smoothed using cubic B-spline method to generate a smoother and more continuous trajectory.

[0078] Cubic B-spline formula:

[0079]

[0080] Among them, P i is the characteristic point of the control curve; F i，k (t) is the K-order B-spline basis function; i is the characteristic point index; k is the order of the B-spline basis function. Substituting k=3 into the cubic B-spline total formula equation, the following formula is obtained to represent the cubic B-spline interpolation basis function:

[0081]

[0082] Step 6: During the journey, the system monitors the environment changes or obstacles in real time. The DWA algorithm quickly replans the local path based on the latest environmental information to ensure the safety and accuracy of navigation. If necessary, the cubic B-spline method is reapplied to smooth the updated path.

[0083] Update path = DWA (current environment information)

[0084] Step 7: When the visually impaired person approaches the destination, the system provides arrival prompts and any necessary final guidance to ensure that the visually impaired person can reach the destination safely and accurately.

[0085] Arrival hint = h (distance from target point)

[0086] If you encounter an obstacle or risk area, simulate the behavior of a hippopotamus escaping from a predator and find a safe path: Use the cubic B-spline method to smooth the new path:

[0087] X i，escape =X i +Random Direction

[0088] Step 8: After confirming that the visually impaired person has safely arrived at the destination, end the navigation service.

[0089] The improvements made to the Hippo algorithm in the present invention are:

[0090] (1) The Tent chaotic mapping iterative formula is introduced to initialize the population sequence of the Hippo algorithm;

[0091] (2) During the exploration and development phases of the Hippo algorithm, the adaptive dynamic adjustment learning factor is used to search for the optimal solution when the position is updated;

[0092] (3) When updating the position, improve the adaptive adjustment based on the number of iterations.

[0093] The improvements are further described below.

[0094] 1. The introduced Tent chaos map iteration formula is defined as:

[0095]

[0096] Among them, r is the chaos parameter, and its value range is [2,4]; x n is the current iteration state, x n+1 is the state of the next iteration.

[0097] 2. In the exploration phase of the Hippopotamus algorithm, when the estimated value of the scattered hippopotamus is reached, the hippopotamus updates its position in the river or pond. The specific formula is:

[0098]

[0099] in, Updated Hippopotamus positions for the exploration phase; Updated front hippo position for exploration phase;X bestis the best position currently found; α is a learning factor used to control the search step size;

[0100] If the improvement of the optimal solution in the current iteration is small, the exploration ability can be enhanced by increasing the learning factor:

[0101] α new =α old ×(1+Δα)

[0102] On the contrary, if the improvement is large, the development capability can be enhanced by reducing the learning factor:

[0103] α new =α old ×(1-Δα)

[0104] Among them, Δα is the preset learning factor adjustment rate; α new is the updated learning factor; α old is the learning factor before updating.

[0105] 3. Hippopotamus defends against predators (exploration phase): Hippopotamus will take defensive actions to avoid being captured by predators. The update formula is:

[0106]

[0107] Among them, predator represents the position of the predator, and β is the defense factor, which is used to simulate the behavior of hippopotamus escaping from predators; Hippopotamus position after the defense phase update.

[0108] 4. Hippopotamus escapes from predators (development phase): The hippopotamus escapes from predators and looks for a safer location. The position update formula is:

[0109]

[0110] Among them, safe_area represents the location of the safe area, and γ is the escape factor, which is used to simulate the behavior of hippopotamus escaping from predators and looking for safe areas; Updated hippo position during the escape phase.

[0111] 5. Adaptive adjustment based on the number of iterations: When updating the position, as the number of iterations increases, the exploration parameter is gradually reduced and the development parameter is increased, so that the search near the optimal solution can be more refined in the later stage of the search:

[0112] α t =α0×e -λt

[0113] β t =β0×(1-e -λt )

[0114] Among them, α t and β t are the learning factor and balance parameter of the tth iteration, λ is the decay rate, and t is the number of iterations.

[0115] The above description is only a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements without departing from the principle of the present invention, and these improvements should also be regarded as within the protection scope of the present invention.

Claims

1. An indoor path planning method based on improved Hippo algorithm and DWA, characterized by: The following steps are involved: S1: The user activates the guide robot dog system through voice commands, the stable gimbal is adjusted to the working state, the sensor is ensured to operate stably, and a path grid map is created with the user's needs as the guide target point; S2: Improve the Hippo algorithm. The improvements are as follows: (1) The Tent chaotic mapping iterative formula is introduced to initialize the population sequence of the Hippo algorithm; (2) During the exploration and development phases of the Hippo algorithm, the adaptive dynamic adjustment learning factor is used to search for the optimal solution when the position is updated; (3) Improve the adaptive adjustment based on the number of iterations when updating the position; S3: Input the path grid map created by S1 into the improved Hippo algorithm and output the global planning trajectory; S4: Use the DWA algorithm to locally refine the global planning trajectory, generate a local collision avoidance trajectory, and obtain the optimal trajectory; S5: Use the B-spline optimization method to smooth the optimal trajectory generated by the DWA algorithm and generate the optimal speed and angular acceleration instructions; S6: When the user approaches the target point, the system provides an arrival prompt to ensure that the visually impaired person can reach the destination safely and ends the navigation service.

2. The indoor path planning method based on the improved Hippo algorithm fused with DWA as claimed in claim 1, characterized in that: The introduced Tent chaotic map iteration formula is defined as: Among them, r is the chaos parameter, and its value range is [2,4]; x n is the current iteration state, x n+1 is the state of the next iteration.

3. The indoor path planning method based on improved Hippo algorithm and DWA as claimed in claim 1, characterized in that: In the exploration phase of the Hippopotamus algorithm, when the estimated value of the scattered hippopotamus is reached, the hippopotamus updates its position in the river or pond. The specific formula is: in, Updated Hippopotamus positions for the exploration phase; Update the front hippo position for the exploration phase; χ best is the best position currently found; α is a learning factor used to control the search step size; If the improvement of the optimal solution in the current iteration is small, the exploration ability can be enhanced by increasing the learning factor: a new =a old ×(1+D) On the contrary, if the improvement is large, the development capability can be enhanced by reducing the learning factor: a new =a old ×(1-D) Among them, Δα is the preset learning factor adjustment rate; α new is the updated learning factor; α old is the learning factor before updating; During the exploration phase, the hippopotamus will take defensive actions to avoid being captured by predators. The update formula is: Among them, predator represents the position of the predator, and β is the defense factor, which is used to simulate the behavior of hippopotamus escaping from predators; Hippopotamus position after the defense phase update.

4. The indoor path planning method based on improved Hippo algorithm and DWA as claimed in claim 1, characterized in that: In the development phase, the hippopotamus escapes from the predator and looks for a safer location. The position update formula is: Among them, safe_area represents the location of the safe area, and γ is the escape factor, which is used to simulate the behavior of hippopotamus escaping from predators and looking for safe areas; Updated hippo position during the escape phase.

5. The indoor path planning method based on improved Hippo algorithm and DWA as claimed in claim 1, characterized in that: When updating the position, as the number of iterations increases, the exploration parameter is gradually reduced and the development parameter is increased so that the search near the optimal solution can be more refined in the later stage of the search. α t ＝α0×e -λ β t ＝β0×(1-e -λt ) Among them, α t and β t are the learning factor and balance parameter of the tth iteration, λ is the decay rate, and t is the number of iterations.

6. The indoor path planning method based on improved Hippo algorithm and DWA as claimed in claim 1, characterized in that: The cubic B-spline curve is used to smooth the path after the redundant nodes are removed. The total formula of the cubic B-spline is: Among them, P i is the characteristic point of the control curve; F i，k (t) is the K-order B-spline basis function; i is the characteristic point index; k is the order of the B-spline basis function. Substituting k=3 into the cubic B-spline total formula equation, the following formula is obtained to represent the cubic B-spline interpolation basis function:

7. A wearable guide robot dog, comprising a robot dog body (3), wherein the robot dog body is provided with a laser radar (1) and a camera (2), and wherein: The robot dog body is also provided with a memory and a processor, the memory stores a computer program, and the processor is used to call and run the computer program stored in the memory to execute the method as described in any one of claims 1 to 6, so that the robot dog body performs navigation services according to the method.

8. The wearable guide robot dog according to claim 7, characterized in that: The invention also comprises a quick-release wearing structure (4), wherein the quick-release wearing structure (4) comprises a right-angle block (41), a wrench (42), a locking tongue (43), a tension spring and a wearing unit (44), wherein the two right-angle blocks (41) are respectively fixed to the robot dog body (3) and the wearing unit (44), wherein each right-angle block (41) is provided with a locking hole (410), wherein the locking tongue (43) is rotatably connected in the locking hole (410), wherein the wrench (42) is fixed to the locking tongue (43), wherein two ends of the tension spring are respectively connected to the wrench (42) and the right-angle block (41) where the wrench is located, wherein the two right-angle blocks (41) are spliced ​​to each other, and wherein the locking tongue (43) on each right-angle block (41) is inserted into the locking hole on the other right-angle block (41).

9. The wearable guide robot dog according to claim 8, characterized in that: A V-shaped hole (411) and a cone head (412) are provided on each right-angle block (41). After the two right-angle blocks (41) are spliced ​​together, the V-shaped holes and the cone heads on the two right-angle blocks are plugged into each other.

10. The wearable guide robot dog according to claim 8, characterized in that: The wearable unit (44) comprises a strap (441) and a pad (442), wherein the strap (441) is connected to the pad (442), wherein a right-angle block (41) is inserted into the strap (441), and an arc pad (443) is provided on the pad (442).

Images