Presentation is loading. Please wait.

Presentation is loading. Please wait.

AmirKabir University Computer Engineering Dept. Sensor Based Mapping and Navigation Robotic Course Presentation Presenter: GholamHossein Deshmeh.

Similar presentations


Presentation on theme: "AmirKabir University Computer Engineering Dept. Sensor Based Mapping and Navigation Robotic Course Presentation Presenter: GholamHossein Deshmeh."— Presentation transcript:

1 AmirKabir University Computer Engineering Dept. Sensor Based Mapping and Navigation Robotic Course Presentation Presenter: GholamHossein Deshmeh

2 Mapping and Navigation Introduction What is Robotic Mapping ? What is Robotic Mapping ? –Building Models of Physical Environments using Robots –Sensors with limitations –Cameras (Vision), Laser, Sonar, GPS

3 Mapping and Navigation Introduction Navigation Navigation –Planning Robot’s Path –Avoiding Obstacles –Fast Re-Planning Mapping + Navigation -> Autonomous Mobile Robots Mapping + Navigation -> Autonomous Mobile Robots

4 Mapping and Navigation Introduction Navigation Mapping Localization

5 Mapping and Navigation Introduction Mapping Mapping –Defines an Environment Model Localization Localization –Probabilistic approach to evaluate the localization on the Environment Model Navigation Navigation –Defines an Optimized Path to the Goal, based on Environment Model

6 Mapping and Navigation Introduction Main Problems Main Problems –Noise –Dimensionality –Planning Exploration –Dynamic Environments

7 Mapping and Navigation Introduction Noise Noise –Errors can Accumulate Dimensionality Dimensionality –3-D Maps can take a high amount of storage space Planning Exploration Planning Exploration –How can a robot explore, using an incomplete map

8 Mapping and Navigation Introduction Dynamic Environments Dynamic Environments –Moving furniture, doors –Moving Cars, People –Must Distinguish between Sensor Noise and Moving Objects

9 Mapping and Navigation Approaches Two decades of research Two decades of research Assume static, small scale environments Assume static, small scale environments Difficulty with dynamic, large scale environments Difficulty with dynamic, large scale environments Simulated vs. Real Life Simulated vs. Real Life

10 Mapping and Navigation Approaches Considerable research in the field of Navigation Considerable research in the field of Navigation Few complete and working systems Few complete and working systems Cost of Stereo Systems Cost of Stereo Systems High Computational Cost High Computational Cost Mostly Laser / Sonar sensors Mostly Laser / Sonar sensors

11 Mapping and Navigation Approaches Kalman Filter Kalman Filter Expectation Minimization (EM) Expectation Minimization (EM) Hybrid Hybrid Occupancy Grids Occupancy Grids

12 Mapping and Navigation Basics A map is a collection of sensor scans, o, and robot positions, s A map is a collection of sensor scans, o, and robot positions, s For every time, t, a new data scan and pose is added to the map: For every time, t, a new data scan and pose is added to the map:

13 Mapping and Navigation Basics Likelihood functions are defined for each available map Likelihood functions are defined for each available map Likelihood function Maximization - obtain the most likely map Likelihood function Maximization - obtain the most likely map Maximization is not possible in Real-Time Maximization is not possible in Real-Time

14 Mapping and Navigation Basics Likelihood Function Maximization Problem. Two Approaches: Likelihood Function Maximization Problem. Two Approaches: –Assume map is correct, add new data (large error growth) –Have the robot stop and calculate after every scan (not real-time)

15 Mapping and Navigation Basics Incremental Localization Incremental Localization –Assume previous map and localizations are accurate –Append new sensor scans to the old map –Localize based on updated map –Can be done in real-time –Fail on cyclic environments as error grows unbounded

16 Mapping and Navigation Basics Incremental Localization (IL) Incremental Localization (IL) –IL never corrects old errors based on new information –Errors can grow unbounded –While traversing a cycle in a map, error growth leads the robot to “get lost” and the map breaks down

17 Mapping and Navigation Basics Expectation Maximization Expectation Maximization –Store scans and position data probabilistically –Search through all possible previous maps (from times 0-t) and find the most likely map –Can handle cyclic environments –Batch algorithms - not real-time

18 Mapping and Navigation Basics Occupancy Grid Mapping: Occupancy Grid Mapping: –The most widely used mapping method –Simple –Many kinds of sensors are acceptable –Adapts well to dynamic Environments

19 Mapping and Navigation Basics Occupancy Grid Mapping – Approach : Occupancy Grid Mapping – Approach : –Divide the Environment into a discrete Grid –Assigns a value to each cell, indicating the Probability that the cell is occupied by an obstacle (Certainty Value, CV) –Initialize all values to 50% ( equal probability that the Location is occupied or Unoccupied ) –Sensor Data -> Uncertainty Regions for an Obstacle -> Increase or Decrease Values

20 Mapping and Navigation Ultrasonic Sensors Conical field of View Conical field of View Returns a measure of the distance to the nearest object within the cone Returns a measure of the distance to the nearest object within the cone Does not specify objects’ Angular location Does not specify objects’ Angular location Objects located on Acoustic Axis are more likely to produce an echo. Objects located on Acoustic Axis are more likely to produce an echo. This is considered in Certainty Grid This is considered in Certainty Grid

21 Mapping and Navigation Ultrasonic Sensors 24 Sensors on a horizontal ring 24 Sensors on a horizontal ring Ring scanning, no need to rotate Ring scanning, no need to rotate A full panorama requires 100 ms to 500 ms A full panorama requires 100 ms to 500 ms All Sensors CANNOT be fired at once All Sensors CANNOT be fired at once Too many senses at the same moment, causes significant CrossTalk Too many senses at the same moment, causes significant CrossTalk

22 Mapping and Navigation Ultrasonic Sensors Mapping Using Sonar Sensors

23 Mapping and Navigation Ultrasonic Sensors CVs are Updated using a heuristic probability function, which considers sensor characteristics. CVs are Updated using a heuristic probability function, which considers sensor characteristics. For sonar sensors, higher values are assigned to Cells in the acoustic axis For sonar sensors, higher values are assigned to Cells in the acoustic axis Calculations must be made for every cell in the sonar sensor’s cone Calculations must be made for every cell in the sonar sensor’s cone

24 Mapping and Navigation Histogramic In-Motion Mapping Improves on Certainty/Occupancy Grid method Improves on Certainty/Occupancy Grid method Certainty Grid method is computationally intensive and can affect real-time robots Certainty Grid method is computationally intensive and can affect real-time robots Histogramic In-Motion Mapping (HIMM) uses a simplification of Certainty Grid Histogramic In-Motion Mapping (HIMM) uses a simplification of Certainty Grid

25 Mapping and Navigation Histogramic In-Motion Mapping Increments only one cell in the HIMM Grid for each sensor reading Increments only one cell in the HIMM Grid for each sensor reading For sonar sensors, this is the cell at the acoustic axis of the sensor For sonar sensors, this is the cell at the acoustic axis of the sensor May seem Oversimplification May seem Oversimplification By sampling each sensor continuously, a probability distribution is obtained By sampling each sensor continuously, a probability distribution is obtained Assumes a Moving robot Assumes a Moving robot

26 Mapping and Navigation Histogramic In-Motion Mapping Less Accurate when the robot is stationary Less Accurate when the robot is stationary Points close to the actual obstacle location, receive a high value Points close to the actual obstacle location, receive a high value Only values of the points on the acoustic axis line are decremented Only values of the points on the acoustic axis line are decremented

27 Mapping and Navigation Histogramic In-Motion Mapping HIMM – Only one cell is updated per reading

28 Mapping and Navigation Histogramic In-Motion Mapping Determining the numerical amount for updating Certainty Values (CV) is an issue Determining the numerical amount for updating Certainty Values (CV) is an issue A too large value would make the robot react to single, possibly false readings A too small value would not build up CVs in time for an avoidance maneuver Suggested values: Suggested values: –Increment : +3 –Decrement : -1

29 Mapping and Navigation Histogramic In-Motion Mapping HIMM - Suggested CVs, Increments and Decrements

30 Mapping and Navigation Histogramic In-Motion Mapping Stationary vs. Mobile Robot Stationary vs. Mobile Robot Stationary : Stationary : –After reaching the maximum set for CV, further readings will be lost –Resulting Cluster will be only ONE Cell

31 Mapping and Navigation Histogramic In-Motion Mapping Squared Certainty Value (SCV) Expresses Confidence Single Reading : could be noise or CrossTalk Multiple Reading : Actual Object Multiple Reading : Actual Object Stronger Response of Navigation Planner (Obstacle Avoidance) when Clusters of SCVs are encountered Stronger Response of Navigation Planner (Obstacle Avoidance) when Clusters of SCVs are encountered Weaker Response for single, unclustered cells Weaker Response for single, unclustered cells

32 Mapping and Navigation Histogramic In-Motion Mapping Obstacle Cluster Strength (OCS): Obstacle Cluster Strength (OCS): –Sum of all SCVs in a cluster ( a group of neighboring cells with CV > 0 ) Helps avoid obstacles soon enough Helps avoid obstacles soon enough Specially Important for Higher speed Vehicles ( V > 0.5 m/s ) Specially Important for Higher speed Vehicles ( V > 0.5 m/s ) Lower OCS for Mobile Robots is a Problem Lower OCS for Mobile Robots is a Problem

33 Mapping and Navigation Histogramic In-Motion Mapping CARMEL Robot, An Example: CARMEL Robot, An Example: –Max Speed: V = 0.78 m/s –D = 100 meters. Minimum distance needed for an obstacle avoidance maneuver –R = 200 meters. Initial Detection of an object –(R – D) / V = 1.28 s –Tp = 160 msec, Sensor firing time –1.28 s / 160 msec = 8 readings total A mapping algorithm must build OCS quickly A mapping algorithm must build OCS quickly Meanwhile must detect erroneous readings Meanwhile must detect erroneous readings Growth Rate Operator (GRO) is Introduced Growth Rate Operator (GRO) is Introduced

34 Mapping and Navigation Histogramic In-Motion Mapping GRO Opr – a: Original CVs, b: Mask, c: Updated CVs

35 Mapping and Navigation Why Vision ? Sonars: slow, not for 3D Mapping Laser-based: expensive and/or slow Vision-based sensors: 1) Good for building 3D models and large scale models. 2) Good for recognition 3) Fast and affordable Speed and Affordability depends on the quality of the stereo system (The more robust the stereo, the slower and/or more expensive it is)

36 Mapping and Navigation Visual Sensor Where uncertainty comes from ? – –From the limitations of the camera – –From the complexity of the environment What are parameters that determine the reliability of data ? – –Feature match error – –Depth calculation error

37 Mapping and Navigation Example I Occupancy Grid Mapping – Original Picture

38 Mapping and Navigation Example I Occupancy Grid Mapping – Disparity Image ( Black : Invalid, Brighter : Closer to Camera )

39 Mapping and Navigation Example I Occupancy Grid Mapping Estimation of Clear, Occupied Regions

40 Conclusion Introduction to Mapping Introduction to Mapping Brief Description and Comparison of Methods Brief Description and Comparison of Methods Occupancy Grid Method Occupancy Grid Method Sensors (Sonar) and Mapping using Occupancy/Certainty/Histogram Grid Sensors (Sonar) and Mapping using Occupancy/Certainty/Histogram Grid Vision and Occupancy Grid Vision and Occupancy Grid

41 References J. Borenstein, Y. Koren. “HISTOGRAMIC IN-MOTION MAPPING FOR MOBILE ROBOT OBSTACLE AVOIDANCE”, IEEE Journal of Robotics and Automation, Vol. 7, No. 4, 1991, pp. 535-539. Chris Urmson, M. Bernardine Dias, “Stereo Vision Based Navigation for Sun-Synchronous Exploration“, Robotics Institute, Carnegie Mellon University Jos´e Santos-Victor, Alexandre Bernardino. “Vision-based Navigation, Environmental Representations and Imaging Geometries” Don Murray, Jim Little. “Using real-time stereo vision for mobile robot navigation”, Computer Science Dept., University of British Columbia

42 Thanks


Download ppt "AmirKabir University Computer Engineering Dept. Sensor Based Mapping and Navigation Robotic Course Presentation Presenter: GholamHossein Deshmeh."

Similar presentations


Ads by Google