Skip to main content

DESKBOT UPDATE 2

Project Overview

Project Goals:
  • Perform Image segmentation and object detection on a cluttered environment.
  • Enable multiple robot behaviours capable of aligning and pushing objects (within a payload).
  • Enable robot to clean surface using coverage path planning.

DeskBot System Workspace:

DeskBot Overview:

DeskBot gives two task option on the Graphic User Interface (GUI). 1) De-clutter, which de-clutters  the table by putting all the objects lying on the table in its respective place and 2) CleanOMatic, This works only when there are no objects lying on the table. The robot is integrated with a 3D printed brush and a defined map helps the robot move around the table to clean the dust particles on the table and dump the dust into an attached bin to the table.

Figure below shows an overview of the process. First, the GUI gives two options 1) De-Clutter, 2) CleanOMatic. Once the power is on the camera is initiated and it perceives the environment. In our case the office table with objects, and hamster robots. When De-clutter has selected the robot from its location plans a path to move the objects to their end location. CleanOMatic cleans the dust on the table with the help of a 3D printed brush integrated with the hamster robot. The robot moves along a definitive path on in the map to clear all the dust on the table and dump them in a dustbin which is attached to one end of the table. 






Components Required:

•Hardware 
•2D Camera 
•Linux Based System 
•Hamster Robots – 2 
•Office Table 
•Miscellaneous (Pens, Markers, Books, Etc)
•Software 
•Robot Operating System 
•Python 
•Tensorflow, Keras, Pytorch 
•OpenCV 
•GUI Interface


Timeline:
•Literature Survey & Proposal - March 12
•Object Detection and Mask R-CNN - March 26
•Coverage Path Planning - April 9
•Path Planning for Object Position Manipulation - April 23

•Final Integration and Final Report - April 30

Comments

Post a Comment

Popular posts from this blog

DESKBOT Update 6

Robot Path Planning for Object Manipulation Over the past several weeks we've shown how the DeskBot system performs Scene segmentation and Object detection. We have also covered the Hamster robot and its role. In this update lets dive into the robot operations.  Once the system is capable of perceiving the scene and detect all the objects present in it, we can enable our robot with multiple behaviors aimed at manipulating an object and pushing it towards a goal. From the Scene Segmentation and Object Detection module we get the following, Positions and centroids of the robot Positions and centroid of the detected unique objects Table dimensions and location of the object holder goal points Euclidian distance from the objects to their respective goal holder Figure below shows a technical overview of how Robot – object manipulation is achieved. Using all the above information from the scene segmentation module we can obtain closed loop vision fe...
Post a Comment
Read more

DESKBOT UPDATE 3

Scene Segmentation and Object Detection The robot has to know its environment before taking an action, a sensor is required is to perceive the environment and know what things exist. In our case, we use a 2D camera to know where and how the objects and the robots are positioned. We use Mask R-CNN to perform instance segmentation and object detection or use YOLO for object detection.  Mask R-CNN is divided into two modules, first, it estimates the regions where the objects can exist on the input image. Second, based on the initial estimation it identifies the class of the object and generates a mask in the pixel level. In the initial step, the RPN (Residual Pooling Network) scans all FPN (Feature Pyramid Network) in a top-bottom approach and estimates where the objects exist on the input image. Once the estimation is done a bounding box is assigned to the anchor (anchors are a set of boxes with predefined locations). RPN helps in the anchor to decide where in the feature...
Post a Comment
Read more

DESKBOT Update 5

YOLO - You Only Look Once   Other methods use classifiers or localizers to perform detection. They apply the model to an image at multiple locations and scales. The high scoring regions of the image are considered detection.        In YOLO the approach is quite different. Neural network is applied for the full image and this network divides the full image into regions and predicts a bounding box or bounding boxes. These bounding boxes are given weights by the predicted probability.        We were supposed to use MaskrCNN to get better accuracy but due to the COVID 19 situation and work from home we do not have a good computing system and GPU resources therefore we have chosen to work with YOLO. YOLO is faster and gives a boundary boxes for the objects detected.        YOLO works on COCO dataset, which is trained on 80 different classes. YOLO is not trained for detecting p...
Post a Comment
Read more