Building outer wall inspection robot and inspection method thereof

Abstract

The invention relates to the technical field of thermal insulation outer wall detection, in particular to a building outer wall inspection robot and an inspection method thereof. The wall climbing mechanisms are arranged on the two sides of the machine shell body, and suction cups are arranged on the faces, making contact with the wall, of the wall climbing mechanisms; the knocking detection mechanism comprises a knocking assembly arranged at one end of the machine shell body, a sound collection assembly close to the knocking end of the knocking assembly and a processor electrically connected with the sound collection assembly. According to the robot, wall climbing operation of the robot is achieved through the suction cups arranged at the bottom of the wall climbing mechanism, the knocking assembly knocks the wall surface, sound is collected and transmitted to the processor, whether the thermal insulation outer wall has defects or not is judged, compared with the prior art, manual work is replaced with the processing mode, and the outer wall inspection efficiency and precision are improved. The invention further requests to protect an inspection method of the inspection robot.

Description

Technical Field

[0001] The present invention relates to the technical field of thermal insulation exterior wall detection, and in particular to a building exterior wall inspection robot and an inspection method thereof. Background Art

[0002] The insulation layer of the building's exterior wall plays an important role in the field of building energy conservation. It is not only directly related to the building's energy efficiency, but also affects the comfort of the living and working environment. The energy consumption of urban heating in winter accounts for a large part of the building's energy consumption, and the impact of energy consumption caused by poor insulation of the building envelope structure is even greater. Therefore, how to improve the insulation and heat insulation performance of the building's exterior, especially the insulation exterior wall, is the key to improving the building's energy efficiency; however, in the current building structure standards, the design service life of the building's insulation system is only 25 years. After the completion of the building, due to temperature changes, grassroots structural factors, improper wall interface treatment and other reasons, the insulation exterior wall is prone to hollowing, cracks, water seepage and other problems, which not only causes huge energy loss, but also may endanger personal safety.

[0003] The existing building exterior wall defect detection methods mainly include manual tapping, infrared thermal imaging, laser ultrasonic vibration, gamma ray or X-ray detection methods. The traditional manual tapping method relies on the experience and judgment of the inspector, which is highly subjective and inaccurate. The high-altitude detection conditions are harsh and cannot be detected for a long time. The infrared thermal imaging method is easily affected by the surface material, surface treatment process, internal structure, and ambient temperature, and has poor stability. Laser ultrasonic vibration and ray detection rely on professional equipment, have high requirements for the detection environment, and have high damage costs.

[0004] In summary, the existing technologies cannot well meet the current demand for intelligent detection of defects in external insulation layers. Summary of the invention

[0005] In view of at least one of the above technical problems, the present invention provides a building exterior wall inspection robot and an inspection method thereof, which adopts a bionic robot to inspect insulation exterior wall defects using AI acoustics to improve the accuracy and efficiency of detection.

[0006] According to a first aspect of the present invention, there is provided a building exterior wall inspection robot, comprising:

[0007] Housing body;

[0008] A wall climbing mechanism is arranged on both sides of the housing body, and a suction cup is provided on the side of the wall climbing mechanism that contacts the wall;

[0009] The knock detection mechanism includes a knock component arranged at one end of the casing body, a sound collection component close to the knocking end of the knock component, and a processor electrically connected to the sound collection component. The bottom of the casing body also has an adsorption mechanism arranged toward the wall.

[0010] In some embodiments of the present invention, the adsorption mechanism is configured to be adsorbed on the wall when the knock detection mechanism is working, and to be separated from the wall when the knock detection mechanism stops working.

[0011] In some embodiments of the present invention, the adsorption mechanism includes an adsorption plate arranged parallel to the bottom of the casing body, a linear driving component connected to the adsorption plate for driving the adsorption plate to move away from or close to the casing body, and an exhaust component connected to the adsorption plate.

[0012] In some embodiments of the present invention, the wall-climbing mechanism is a bionic mechanical foot, the suction cup is fixed to the sole of the bionic mechanical foot, and the bionic mechanical foot further has a roller mechanism on a side away from the suction cup.

[0013] In some embodiments of the present invention, the roller mechanism includes a rotating wheel and a rotating driving member drivingly connected to the rotating wheel, and the housing body also has a switching driving member connected to the bionic mechanical foot.

[0014] In some embodiments of the present invention, the knocking component includes a knocking vibration hammer that can move relatively toward or away from the wall and a knocking drive connected to the knocking vibration hammer; the sound collection component includes a microphone and a sound insulation shell that surrounds the end of the knocking vibration hammer and is arranged toward the microphone opening.

[0015] In some embodiments of the present invention, the housing body also has a video acquisition device and / or a laser radar for obstacle identification and path planning, and the video acquisition device and / or the laser radar are electrically connected to the processor.

[0016] According to a second aspect of the present invention, there is also provided an inspection method of the building exterior wall inspection robot, comprising the following steps:

[0017] Drive the robot to the set location on the wall;

[0018] Driving the knocking component to knock the wall;

[0019] Collect the knocking sound and transmit it to the processor;

[0020] The processor uses a feature extraction method based on wavelet transform of time-frequency domain analysis of sound signals to obtain a time-frequency diagram of the sound signal, and inputs the time-frequency diagram into a trained debonding defect classification algorithm to analyze whether a defect exists.

[0021] In some embodiments of the present invention, when the knocking component is driven to knock on the wall, the adsorption mechanism is controlled to be adsorbed on the wall.

[0022] In some embodiments of the present invention, when driving the robot to move, a grid method is used for environment modeling, and an inner spiral method is used for path planning and obstacle avoidance.

[0023] The beneficial effects of the present invention are as follows: the present invention realizes the wall-climbing operation of the robot by means of a suction cup arranged at the bottom of the wall-climbing mechanism, and knocks on the wall through a knocking component, collects sound and transmits it to a processor to judge whether there are defects in the thermal insulation exterior wall. Compared with the prior art, the above processing method replaces manual labor, improves the efficiency and accuracy of exterior wall inspections, and through the setting of the adsorption mechanism, it can improve the stability during knocking, avoid accidental falling of the robot, and reduce the detection cost in disguise. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0025] Figure 1 This is a schematic diagram of the structure of a building exterior wall inspection robot in an embodiment of the present invention;

[0026] Figure 2 This is a schematic structural diagram of the building exterior wall inspection robot from another perspective in an embodiment of the present invention;

[0027] Figure 3 This is a structural schematic diagram of the building exterior wall inspection robot from another perspective in an embodiment of the present invention;

[0028] Figure 4 In the embodiment of the present invention Figure 1 A local enlarged view of point A in FIG.

[0029] Figure 5 A schematic diagram of the structure of a knocking component and a sound collection component in an embodiment of the present invention;

[0030] Figure 6 This is a schematic diagram of the structure of the building exterior wall inspection robot when knocking in an embodiment of the present invention;

[0031] Figure 7 This is a schematic diagram of the structure of the building exterior wall inspection robot when crawling in an embodiment of the present invention;

[0032] Figure 8The present invention is a flowchart of the inspection method of the building exterior wall inspection robot in an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0034] It should be noted that when an element is referred to as being "fixed to" another element, it may be directly on the other element or there may be a central element. When an element is considered to be "connected to" another element, it may be directly connected to the other element or there may be a central element at the same time. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only and do not represent the only implementation method.

[0035] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0036] like Figures 1 to 7 The building exterior wall inspection robot shown in the figure includes a housing body 1, a wall climbing mechanism 2 and a knock detection mechanism 3, as shown in FIG. Figure 1 and Figure 2As shown in , the casing body 1 can be formed by assembling the shell to play a supporting and fixing role. A circuit board and a driving structure for power supply and driving can be set in the casing body 1; in an embodiment of the present invention, the wall climbing mechanism 2 is arranged on both sides of the casing body 1, and the wall climbing mechanism 2 has a suction cup 21 on the side in contact with the wall; in some embodiments of the present invention, the wall climbing mechanism 2 adopts a three-degree-of-freedom mechanical foot, and through the setting of the suction cup 21, it can generate an adsorption force on the wall, and crawling is achieved through the movement of the joint. The mechanical foot is a prior art in this field and will not be described in detail here. In an embodiment of the present invention, the knocking detection mechanism 3 includes a knocking component 31 arranged at one end of the casing body 1, a sound collection component 32 near the knocking end of the knocking component 31, and a processor 33 electrically connected to the sound collection component 32. In addition, in an embodiment of the present invention, the bottom of the casing body 1 also has an adsorption mechanism 4 arranged toward the wall. Through the setting of the adsorption mechanism 4, the robot can be adsorbed on the wall when knocking, thereby improving the reliability of detection. The specific detection principle is that when the outer insulation layer has bonding defects such as hollowing of the exterior wall finishing layer, defects in the bonding layer, and hollowing of the insulation layer, when the insulation exterior wall is struck, due to insufficient bonding integrity, different vibration modes will be presented, thereby generating different vibration frequencies. In an embodiment of the present invention, the processor 33 uses a NI sound processing device and an AI smart tablet to analyze and identify the sound. The specific analysis method is described in detail below.

[0037] In the above embodiment, the robot wall climbing operation is realized by the suction cup 21 provided at the bottom of the wall climbing mechanism 2, and the wall is knocked by the knocking component 31, and the sound is collected and transmitted to the processor 33 to determine whether there are defects in the insulation exterior wall. Compared with the prior art, the above processing method replaces manual labor and improves the efficiency and accuracy of exterior wall inspection.

[0038] On the basis of the above embodiments, in order to avoid the risk of the robot being unstable or falling when knocking, in some embodiments of the present invention, the adsorption mechanism 4 is configured to be adsorbed on the wall when the knock detection mechanism 3 is working, and to be separated from the wall when the knock detection mechanism 3 stops working. Figure 3In some embodiments of the present invention, the adsorption mechanism 4 includes an adsorption plate 41 arranged parallel to the bottom of the housing body 1, a linear drive member 42 connected to the adsorption plate 41 for driving the adsorption plate 41 to move away from or close to the housing body 1, and an exhaust member 43 connected to the adsorption plate 41. The exhaust member 43 here can be an air pump, and the linear drive member 42 can be an electric push rod or other linear drive structure. Before the actual knocking, the adsorption plate 41 is attached to the wall through the linear drive member 42, and then the air is sucked through the exhaust member 43 so that the hole position on the adsorption plate 41 generates negative pressure and then generates suction with the wall, thereby avoiding vibration of the robot when knocking, improving the accuracy of sound collection and improving the reliability of the robot. Of course, it should be pointed out here that the adsorption mechanism 4 has a variety of structural forms, for example, the structural form of an adsorption disk can also be adopted. Those skilled in the art can make conventional choices based on the concept of the present invention, but all fall within the scope of protection of the present invention.

[0039] In the embodiment of the present invention, the wall climbing mechanism 2 is a bionic mechanical foot 22, the suction cup 21 is fixed to the sole of the bionic mechanical foot 22, and the side of the bionic mechanical foot 22 away from the suction cup 21 also has a roller mechanism 23. It should be pointed out here that the bionic mechanical foot 22 is a prior art in the field, for example, the mechanical foot can be assembled by angle brackets, steering gears, U-shaped brackets, etc., the longitudinal movement of the suction cup 21 is achieved by one steering gear, and the movement process such as forward or backward is achieved by another steering gear. For details, please refer to Figure 7 In the embodiment of the present invention, four bionic mechanical feet 22 are provided. The four bionic mechanical feet 22 cooperate with each other to realize the flexible movement of the body, improve the maneuverability of the body, and are conducive to dealing with gaps or protrusions on the wall of the building. When climbing the wall, there is an adsorption force generated by the suction cup 21 at the bottom, and the movement and steering of the robot are realized through the lateral joint and the forward joint. In the embodiment of the present invention, in addition to crawling, the roller mechanism 23 can also be used to realize rapid movement on flat ground. For details, please refer to Figure 4 The roller mechanism 23 includes a rotating wheel 23a and a rotating driving member 23b drivingly connected to the rotating wheel 23a, and the housing body 1 also has a switching driving member 24 connected to the bionic mechanical foot 22. The switching driving member 24 can be a steering gear, which turns the mechanical foot 180 degrees so that the roller mechanism 23 faces downward and contacts the ground, and the synchronous rotation of the roller is achieved by rotating the driving member 23b, thereby achieving the purpose of rapid movement on flat ground.

[0040] In the embodiment of the present invention, the specific structure of the knocking component 31 is as follows: Figure 5As shown in the figure, the knocking assembly 31 includes a knocking vibration hammer 31a that can move relatively in the direction close to or away from the wall and a knocking driving member 31b connected to the knocking vibration hammer 31a; of course, it should be pointed out here that the knocking driving member 31b has a variety of structural forms, which can be Figure 5 As shown in the figure, the blade drives the percussion vibration hammer 31a to move up and down by the rotation of the rotating cylinder, and it can also be realized by adopting a structure such as a piezoelectric driver or an electromagnetic coil. The sound collection component 32 includes a microphone 32a and a sound insulation shell 32b surrounding the end of the percussion vibration hammer 31a and opening toward the microphone 32a. By setting the sound insulation shell 32b, the interference of external noise can be reduced, and the accuracy and reliability of sound collection can be improved.

[0041] In the embodiment of the present invention, the movement of the robot can be controlled by remote control, for example, by wireless communication modules such as WiFi / 4G / 5G to realize communication between the control center and the robot, or by autonomous path planning. As shown in the figure, the housing body 1 also has a video acquisition device 5 and / or a laser radar 6 for obstacle identification and path planning, and the video acquisition device 5 and / or the laser radar 6 are electrically connected to the processor 33. With the help of the video acquisition device 5 or the laser radar 6, or by using the video acquisition device 5 and the laser radar 6 at the same time, the monitoring of the detection environment and the identification of the obstacle type can be realized, thereby further improving the detection efficiency. The specific path planning and obstacle identification methods will be described in detail below.

[0042] In an embodiment of the present invention, an inspection method of the above-mentioned building exterior wall inspection robot is also provided, such as Figure 8 As shown in , the following steps are included:

[0043] S10: driving the robot to a set location on the wall; the set location here may be achieved by remote control, or may reach the set location along a planned path.

[0044] S20: driving the knocking component 31 to knock on the wall;

[0045] S30: collecting knocking sounds and transmitting them to the processor 33;

[0046] S40: The processor 33 uses a feature extraction method based on wavelet transform of time-frequency domain analysis of sound signals to obtain a time-frequency diagram of the sound signal, and inputs the time-frequency diagram into a trained debonding defect classification algorithm to analyze whether there is a defect.

[0047] In the prior art, acoustic signal detection mostly uses the time-frequency analysis method of short-time Fourier transform. In the short-time Fourier transform, the length of the window determines the time resolution and frequency resolution of the spectrum. The longer the window length, the higher the frequency resolution after the Fourier transform, and the worse the time resolution, which leads to the inability to achieve both the time resolution and the frequency resolution of the acoustic signal. In an embodiment of the present invention, in order to avoid the above problems, a feature extraction method of wavelet transform based on the time-frequency domain analysis of the sound signal is adopted to obtain the time-frequency diagram of the acoustic signal. The wavelet time-frequency diagram represents the component intensity of a certain time and a certain frequency through color, contains more original information, and is more comprehensive than the traditional characterization method.

[0048] In an embodiment of the present invention, after converting the sound signal into an image signal, the recognition of defects is realized by a deep learning algorithm. Specifically, for signal training, recognition and classification, a method combining a convolutional neural network based on image recognition and a support vector machine (CNN-SVM) in deep learning is used to construct a debonding defect classifier. In this way, the time-frequency graph is recognized and classified, the category of the sound signal is accurately judged, and the defect of the insulation wall is determined. After obtaining the wavelet time-frequency graph, a convolutional neural network (CNN) is used to perform deep learning on the image data, extract image features, and automatically extract features through operations such as convolution and pooling. After processing and training by CNN, the image, a high-dimensional data with a spatial structure, is converted into a low-dimensional feature vector. A support vector machine (SVM) is selected for classification training, and SVM has good data generalization ability and efficient data processing ability. In an embodiment of the present invention, four groups of knocking experiments under different debonding conditions are designed to verify the accuracy of the algorithm, and the sound signals collected in nine rectangular areas are trained and recognized respectively, and the classification accuracy meets the design requirements. Through the setting of the above method, not only the efficiency of the thermal insulation exterior wall inspection is improved, but also the accuracy of the inspection is improved.

[0049] In the embodiment of the present invention, when the knocking assembly 31 is driven to knock the wall, the adsorption mechanism 4 is controlled to be adsorbed on the wall. This part has been mentioned above and will not be repeated here. Those skilled in the art can understand it in combination with the above.

[0050] In an embodiment of the present invention, when driving the robot to move, the grid method is used for environmental modeling, and the inner spiral method is used for path planning and obstacle avoidance. Specifically, the robot is equipped with a laser radar sensor and a camera vision sensor. The laser radar sensor realizes the obstacle ranging function, and the camera vision sensor realizes environmental monitoring and obstacle type recognition. The two together assist in realizing intelligent inspection, and at the same time, personnel can understand environmental information based on camera information and intervene remotely when necessary. During the operation, the AI ​​camera at the front end of the inspection robot is first used to scan the wall to obtain wall information, mark the obstacles appearing on the wall, reflect them on the computer, and present a real-time picture of the wall. At the same time, the inspection robot uses its own laser radar during the work process to locate its position, and when emergencies such as window opening occur, it is fed back to the computer, and obstacles are supplemented on the wall map simulated by the AI ​​camera. Finally, the data collected by the dual sensors are used for data fusion to plan the best path for the inspection robot.

[0051] Based on laser radar technology, radar sensors are used to obtain the surrounding environment data of the inspection robot and accurately determine the distance between the robot and environmental objects. Laser radar uses laser light waves as carriers, emits pulsed laser beams, processes the echoes reflected or scattered by the target, and then obtains the distance information of the target in conjunction with a certain imaging method. In an embodiment of the present invention, a MEMS-based scanning method is adopted, and a MEMS mirror is used to realize scanning in the vertical direction, and a laser pulse (repetition frequency is 200kHz, wavelength is 1550nm) and a unit InGaAs detector are used in conjunction with high repetition frequency emission. Based on OpenCV, computer vision and deep learning technology are used to identify and analyze the characteristic elements in the video image. The obstacle type identified by the visual sensor is a three-dimensional obstacle, such as a window frame, a glass fixing frame, and an open window. Then the video image is marked and classified to generate an understanding and reasoning of the content of the video image. The specific implementation process is: video stream → frame-by-frame decoding → feature extraction → neural network training → classification recognition → feedback loop.

[0052] In an embodiment of the present invention, the control of the intelligent navigation method includes two key technical steps: model construction and inner spiral path planning algorithm. Specifically, regarding model construction: for the plane model construction of the thermal insulation exterior wall, the grid method is used. The grid method is to use the same size and size grid to comprehensively divide the robot workspace information path environment, and then complete the entire path environment planning and construction by recording the environmental space information of its own location. Using the grid method for environmental modeling, the following four points are considered: ① wall inspection area; ② formulate specific paths and grid map sizes according to the path planning method; ③ grid size in the environmental map; ④ computer computing power and reaction ability. The wall is regarded as a regular rectangle, the size of the inspection robot is used as the basic unit, and the wall is divided into several non-overlapping cells according to the range of detection required. Each cell represents a free area or an obstacle area. The robot can move freely in the free area, and the robot cannot pass through the obstacle area. The white grid represents the free area, and the black grid represents the obstacle. Then, a matrix is ​​established on MATLAB, with the black grid marked 1 and the white grid marked 0, and its wall model is constructed.

[0053] The specific algorithm for inner spiral path planning is to use the inner spiral path planning algorithm of walking along the edge to realize robot obstacle avoidance detection based on the grid map. The robot moves along the boundary of the wall or obstacle to complete the inspection, so as to achieve the maximum coverage and minimum repetition rate of the robot inspection.

[0054] During the obstacle avoidance process, the following four situations may be encountered: ① Most obstacles are irregular in shape and are all treated as rectangles, that is, the envelope method is used to make the rectangular boundaries of irregular obstacles; ② Obstacles placed along the edge of the wall are usually regarded as boundaries, and the inspection robot walks along the edge of the obstacle when working; ③ Obstacles set on non-edges can be bypassed; ④ The inspection robot constantly adjusts the angle, speed, and direction during walking. After each work, it goes through 4 different directions (up, down, left, and right). In obstacle avoidance path planning, the robot needs to detect obstacles and process data while working. Among them, the boundary can be regarded as an obstacle. Before walking, the robot searches to the right for obstacles. If there are obstacles, it avoids obstacles along the edge of the obstacle according to the inner spiral path planning algorithm of walking along the edge, and reselects the direction of movement. If the obstacle does not exist, it continues to walk in the original direction.

[0055] Those skilled in the art should understand that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A building exterior wall inspection robot, characterized in that: include: Casing body; A wall climbing mechanism is arranged on both sides of the housing body, and a suction cup is provided on the side of the wall climbing mechanism that contacts the wall; The knock detection mechanism includes a knock component arranged at one end of the casing body, a sound collection component close to the knocking end of the knock component, and a processor electrically connected to the sound collection component. The bottom of the casing body also has an adsorption mechanism arranged toward the wall.

2. The building exterior wall inspection robot according to claim 1, characterized in that: The adsorption mechanism is configured to be adsorbed on the wall surface when the knock detection mechanism is working, and to be separated from the wall surface when the knock detection mechanism stops working.

3. The building exterior wall inspection robot according to claim 2, characterized in that: The adsorption mechanism includes an adsorption plate arranged parallel to the bottom of the casing body, a linear driving member connected to the adsorption plate for driving the adsorption plate to move away from or close to the casing body, and an exhaust member connected to the adsorption plate.

4. The building exterior wall inspection robot according to claim 1, characterized in that: The wall climbing mechanism is a bionic mechanical foot, the suction cup is fixed on the sole of the bionic mechanical foot, and the side of the bionic mechanical foot away from the suction cup is also provided with a roller mechanism.

5. The building exterior wall inspection robot according to claim 4, characterized in that: The roller mechanism comprises a rotating wheel and a rotating driving member drivingly connected to the rotating wheel, and the housing body also comprises a switching driving member connected to the bionic mechanical foot.

6. The building exterior wall inspection robot according to claim 1, characterized in that: The knocking component includes a knocking vibration hammer that can move relatively toward or away from the wall and a knocking drive connected to the knocking vibration hammer; the sound collection component includes a microphone and a sound insulation shell that surrounds the end of the knocking vibration hammer and is arranged toward the microphone opening.

7. The building exterior wall inspection robot according to claim 1, characterized in that: The housing body also has a video acquisition device and / or a laser radar for obstacle identification and path planning, and the video acquisition device and / or the laser radar are electrically connected to the processor.

8. An inspection method for a building exterior wall inspection robot according to any one of claims 1 to 7, characterized in that: The following steps are involved: Drive the robot to the set location on the wall; Driving the knocking component to knock the wall; Collect the knocking sound and transmit it to the processor; The processor uses a feature extraction method based on wavelet transform of time-frequency domain analysis of sound signals to obtain a time-frequency diagram of the sound signal, and inputs the time-frequency diagram into a trained debonding defect classification algorithm to analyze whether a defect exists.

9. The inspection method of the building exterior wall inspection robot according to claim 8, characterized in that: When the knocking component is driven to knock the wall, the adsorption mechanism is controlled to be adsorbed on the wall.

10. The inspection method of the building exterior wall inspection robot according to claim 8, characterized in that: When driving the robot to move, the grid method is used for environment modeling, and the inner spiral method is used for path planning and obstacle avoidance.

Images