Presentation is loading. Please wait.

Presentation is loading. Please wait.

ROBOTIC VISION CHAPTER 5: NAVIGATION. CONTENTS INTRODUCTION OF ROBOT NAVIGATION SYSTEM. VISUAL GUIDED ROBOT APPLICATION: - LAND BASED NAVIGATION - MAP.

Similar presentations


Presentation on theme: "ROBOTIC VISION CHAPTER 5: NAVIGATION. CONTENTS INTRODUCTION OF ROBOT NAVIGATION SYSTEM. VISUAL GUIDED ROBOT APPLICATION: - LAND BASED NAVIGATION - MAP."— Presentation transcript:

1 ROBOTIC VISION CHAPTER 5: NAVIGATION

2 CONTENTS INTRODUCTION OF ROBOT NAVIGATION SYSTEM. VISUAL GUIDED ROBOT APPLICATION: - LAND BASED NAVIGATION - MAP BASED NAVIGATION - VISION BASED NAVIGATION

3 CONTENTS INTRODUCTION TO PATH PLANNING PATH PLANNING ALGORITHM SEARCH - A* METHOD

4 INTRODUCTION Navigation is nothing more than plotting an efficient route from point A to point B. Robot navigation includes just two things: the ability to move and a means to determine whether or not the goal has been reached. The problem is how we can finding the most efficient way to reach a destination.

5 INTRODUCTION Navigation is the process of moving the robot to a goal location (or series of locations) in the environment, steering around obstacles along the way. This is accomplished via operations called Direct Motion commands, such as moving forward a given distance, turning to a given angle, or changing (accelerating or decelerating) to a given velocity.

6 INTRODUCTION Robot navigation can be defined as the combination of three basic activities: - Map building - Localization - Path planning

7 MAP BUILDING The process of constructing a map from sensor readings taken at different robot locations. The correct treatment of sensor data and the reliable localization of the robot are fundamental in the map-building process.

8 LOCALIZATION The process of getting the actual robot’s location from sensor readings and the most recent map. An accurate map and reliable sensors are crucial to achieving good localization.

9 PATH PLANNING The process of generating a feasible and safe trajectory from the current robot location to a goal based on the current map. In this case, it is also very important to have an accurate map and a reliable localization.

10 INTRODUCTION In any navigation scheme, it is the desire to reach a destination without getting lost or crashing into anything. The navigation problem is to find a path from start (S) to goal (G) and traverse it without collision. The relationship between the three subtasks mapping and modeling of the environment; path planning and selection; path traversal and collision avoidance  into which the navigation problem is decomposed, is shown in Fig in next slide.

11

12 LAND BASED NAVIGATION Divided into 2 types: - Natural landmark navigation - Artificial landmark navigation Landmarks are distinct features that a robot can recognize from its sensory input. Landmarks can be geometric shapes such as rectangles, lines, circles, etc.

13 LAND BASED NAVIGATION Landmarks are carefully chosen to be easily identifiable; for example, there must be sufficient contrast to the background. Before a robot can use landmarks for navigation, the characteristics of the landmark must be known and stored in the robot's memory.

14 LAND BASED NAVIGATION

15 NATURAL LANDMARK NAVIGATION A natural landmark positioning system has the following basic components: - A sensor for detecting landmarks and contrasting them against their background. - A method for matching observed features with a map of known landmarks. - A method of computing location and localization errors from the matches.

16 NATURAL LANDMARKS NAVIGATION Natural landmarks work best in highly structured environments such as corridors, manufacturing floors, or hospitals. natural landmarks define as those objects or features that are already in the environment and have a function other than robot navigation. Natural landmarks require no modifications to the environment.

17

18 ARTIFICIAL LANDMARK NAVIGATION artificial landmarks define as specially designed objects or markers that need to be placed in the environment with the sole purpose of enabling robot navigation. Detection is much easier with artificial landmarks, which are designed for optimal contrast. In addition, the exact size and shape of artificial landmarks are known in advance. Artificial landmarks are inexpensive and can have additional information encoded on them.

19 MAP BASED NAVIGATION Map-based positioning (also known as "map matching"), is a technique in which the robot uses its sensors to create a map of its local environment. This local map is then compared to the global map previously stored in memory. If a match is found then the robot can compute its actual position and orientation in the environment. The pre stored map can be a CAD model of the environment, or it can be constructed from prior sensor data.

20 MAP BASED NAVIGATION The main advantages of map-based positioning are given below: 1. It uses naturally the naturally occurring structure of typical indoor environments to derive position information without modifying the environment. 2. It can be used to generate an updated map of the environment. Environment maps are important for other mobile robot tasks, such as global path planning. 3. It allows a robot to learn about a new environment and to improve positioning accuracy through exploration.

21 MAP BASED NAVIGATION Disadvantages of map-based positioning arise because it requires that: 1. There be enough stationary, easily distinguishable features that can be used for matching. 2. The sensor map be accurate enough (depending on the task at hand) to be useful. 3. A significant amount of sensing and processing power is available.

22 VISION BASED NAVIGATION Vision based positioning or localization uses the same basic principles of landmark-based and map-based positioning but relies on optical sensors rather than ultrasound, dead- reckoning and inertial sensors.

23 VISION BASED NAVIGATION The advantage of these type of sensors lies in their ability to directly provide distance information needed for collision avoidance. They have an important drawback in that only vertical structures (ie. mainly the shape of the free space surrounding the robot) can be recognised.

24 VISION BASED NAVIGATION Visual sensing can provide the robot with an incredible amount of information about its environment. The most common optical sensors include laser-based range finders and photometric cameras using CCD arrays.

25 EXAMPLE: Navigation of indoor mobile robots The landmark model is designed to have a three-dimensional structure consisting of a multi-colored planar pattern. Experimental results show that the proposed landmark model is effective.

26 Figure shows the appearance of the proposed landmark. The color pattern is composed of two vertically neighboring color patches. The color pattern makes the angle of 45o with respect to two supporting planes at the right angle.

27 Seven landmarks is set at the predetermined positions in laboratory and corridor then the robot estimated its position and corrected the direction of movement along the round path.

28

29

30 INTRODUCTION OF PATH PLANNING Now we want to know how to plan a path for the robot to move through space to a goal position. We want to – Plan a path through the workspaces constrained by borders and obstacles – Find the best path from the set of all possible paths to the goal

31

32 Given an initial configuration start position point and a goal configuration (goal position point), the purpose of path planning to generate the best continuous path of legal configurations between these two points, if such a path exists.

33 PATH PLANNING ALGORITHM Two major families of robot path planning algorithm are cell decomposition and skeletonization. There are many various version of path planning algorithm. All of these version are based on these 2 major families.

34 PATH PLANNING ALGORITHM The following are the various version of path planning algorithm: - Potential methods - Visibility graphs - Voronoi diagrams - Cell decomposition - Wavefront planning - Distance Transform

35 VISIBILITY GRAPH Given polygonal obstacles, shortest path shaves corners of obstacles Algorithm: – Build visibility graph linking all non blocked vertices, start and goal – Search graph for shortest path to goal. Not often used for path planning for mobile robots.

36 VISIBILITY GRAPH EXAMPLE

37 VORONOI GRAPH Alternative approach: build a graph representing safest paths throughout map Algorithm: – Start with C-space map – Draw equidistant lines from obstacles and borders – Add lines to q_goal and q_start – Search graph for shortest path from start to goal

38 VORONOI GRAPH Requires complex algorithms to find Voronoi graph Can be very complex (and slow) in higher dimensions. Based on skeletonization family

39 VORONOI GRAPH EXAMPLE

40 CELL DECOMPOSITION Divide free space into convex polygons and plan between legal cells Algorithm: – Start with free-space map – Divide map into polygons or cells – Mark cells containing obstacles as occupied – Search for path to goal using unoccupied cells

41 CELL DECOMPOSITION Multiple ways to sub-divide all possible robot space. example: Grids Resolution is a limiting issue – Too fine a grid leads to long search time – Too coarse a grid misses paths Requires a search algorithm E.g. A*

42

43 Cell Decomposition Example q_start q_goal Occupied Free

44 8-Connected Cell Decomposition q_start q_goal Occupied Free

45 DISTANCE TRANSFORM Under skeletonization path planning family. Not suitable for big area. Maximal clearance path

46 DISTANCE TRANSFORM Have 3 different distance measure method: - Euclidean distance transform - City block distance transform - Chessboard distance transform - Chamfer metrics DT The Euclidean DT has traditionally not been very popular because of the algorithmic and computational complexity of its implementation. city-block, chessboard and chamfer metrics have enjoyed popular usage because they are easily and efficiently implemented.

47 DISTANCE TRANSFORM

48 Raster-scanning The image is scanned row-by-row or column-by-column in a predefined order. Distance information (scalars or vectors) are propagated using a small neighbourhood.

49 COMPUTE DISTANCE TRANSFORM One of the method to compute distance transform by using two pass sequential algorithm. Originally proposed in 1966 by Rosenfeld and Pfalz for the city-block and chessboard metrics. Generalized in 1984 by Borgefors for chamfer metrics and the hexagonal metric.

50 COMPUTE DISTANCE TRANSFORM The algorithm consists of a forward pass (in raster order) through the image followed by a backward pass (in anti-raster order). Distance is propagated in each pass with respect to a propagation mask.

51 COMPUTE DISTANCE TRANSFORM Masks on the left are for the square grid, and the ones on the right are for the hexagonal grid. Top masks are for the forward pass and the bottom masks are for the backward pass. The centre of each mask is the zero value.

52 Common metrics used in the two-pass algorithm city-block metric: a = 1 and b = 1 chessboard metric: a = 1 and b = 1 3-4-chamfer metric: a = 3 and b = 4 5-7-chamfer metric: a = 5 and b = 7

53

54

55 DISTANCE TRANSFORM EXAMPLE

56

57

58 SEARCH Once you have your graph you need to search for the best path -Your graph can be based on cell decomposition, visibility graph, voronoi diagram, quadtree and etc Several search methods can be used: - A* - Other options include random search, depth first search, breadth first search, etc. Good search techniques are important for a variety of reasons

59 A* SEARCH METHOD Navigating a terrain and finding the shortest path to a Goal location is one of the fundamental problems in path planning. While there are many approaches to this problem, one of the most common and widely known is the A Star search.

60 Often used to search for the lowest cost path from the start point to the goal position location in a graph of cells/voronoi/visibility/quadtree.

61 COMPUTE A* SEARCH METHOD Define g(n)  Cost to reach node n from start Greedy search by following child with lowest f(n)  f(n) = g(n) + h(n) Provided h(n) is admissible, A* will be complete and optimal  Performance will depend on how good h(n) is Admissibility:  h(n) must never overestimate the cost

62 A* Example 8 7 6 9 S 6 5 876 4 3 2 5 5 4 3 6 4 5 6 5 3 4 5 2 3 43 G 1 2 1 1 2 3 2 h(n) Assume a 4-Connected Graph

63 A* Example 8+1 7 6 9 S 6+1 5 8+176 4 3 2 5 5 4 3 6 4 5 6 5 3 4 5 2 3 43 G 1 2 1 1 2 3 2 Queue of children to consider f(n)= h(n)+g(n)

64 A* Example 8+1 7+2 6 9 S 1(7) 5+2 8+176 4 3 2 5 5 4 3 6 4 5 6 5 3 4 5 2 3 43 G 1 2 1 1 2 3 2 h(n) g(n)

65 A* Example 8+1 7+2 6+3 9 S 1(7) 2(7) 8+176 4 3 2 5 5 4 3 6 4 5 6 5 3 4 5 2 3 43 G 1 2 1 1 2 3 2

66 A* Example 8+1 7+2 3(9) 9 S 1(7) 2(7) 8+176 4 3 2 5 5 4 3 6 4 5 6 5+4 3 4 5 2 3 43 G 1 2 1 1 2 3 2

67 A* Example 8+1 7+2 3(9) 9 S 1(7) 2(7) 8+1 76 4 3 2 5 5 4 3 6 4+5 5 6 4(9) 3 4 5 2 3 43 G 1 2 1 1 2 3 2

68 A* Example 8+1 7+2 3(9) 9 S 1(7) 2(7) 8+176 4 3 2 5 5 4 3 6 5(9) 5+6 6 4(9) 3+6 4 5 2 3 43 G 1 2 1 1 2 3 2

69 A* Example 8+1 7+2 3(9) 9 S 1(7) 2(7) 8+176 4 3 2 5 5 4 3 6 5(9) 5+6 6 4(9) 6(9) 4+7 5 2+7 3 43 G 1 2 1 1 2 3 2

70 A* Example 8+1 7+2 3(9) 9 S 1(7) 2(7) 8+1 76 4 3 2 5 5 4 3 6 5(9) 5+6 6 4(9) 6(9) 4+7 5 7(9) 3+8 43 G 1 2 1 1 2 3 2 An interesting thing just happened!!!

71 A* Example 8+1 7+2 3(9) 9+2 S 1(7) 2(7) 1(9)7+26 4 3 2 5 5 4 3 6 5(9) 5+6 6 4(9) 6(9) 4+7 5 7(9) 3+8 43 G 1 2 1 1 2 3 2

72 A* Example 8+1 7+2 3(9) 9+2 S 1(7) 2(7) 1(9)2(9)6+3 4 3 2 5 5 4 3 6 5(9) 5+6 6 4(9) 6(9) 4+7 5 7(9) 3+8 43 G 1 2 1 1 2 3 2

73 A* Example 8+1 7+2 3(9) 9+2 S 1(7) 2(7) 1(9)2(9)3(9) 5(9) 6(9) 7(9) 4(9) 5+6 4+7 3+8 6+5 5(9) 5+6 6 4(9) 6(9) 4+7 5 7(9) 3+8 43 G 1 2 8(9) 1 2 3 2+9

74 A* Example 8+1 7+2 3(9) 9+2 S 1(7) 2(7) 1(9)2(9)3(9) 5(9) 6(9) 7(9) 4(9) 5+6 4+7 3+8 6+5 5(9) 5+6 6 4(9) 6(9) 4+7 5 7(9) 3+8 43 G 1 2 8(9) 1 2 3 2+9 Path Length = 9 steps!!

75 8 7 3 9 S 1 2 876 4 3 2 5 5 4 3 6 5 5 6 4 6 4 5 7 8 910 G 12 11 1 1 2 3 2 Path length=13 steps! 8 7 6 9 S 6 5 123 5 6 7 4 5 4 3 6 4 5 6 5 3 4 5 2 3 43 G 1 2 8 1 2 3 2 A-star (A*) Search Path length=9 steps!

76


Download ppt "ROBOTIC VISION CHAPTER 5: NAVIGATION. CONTENTS INTRODUCTION OF ROBOT NAVIGATION SYSTEM. VISUAL GUIDED ROBOT APPLICATION: - LAND BASED NAVIGATION - MAP."

Similar presentations


Ads by Google