jarcand / arisecar Goto Github PK

The direction of the fog is unknown

Need to implement a simple proof-of-concept program for the IMU once it is 
installed.

Need to implement a simple proof-of-concept program for the rotary encoders 
once they are installed on the wheels.

While power saving is important for a battery powered robot, the Win7 computer 
is automatically hibernating which is uncontrolled and could cause problems for 
the demo.

So, before the demo, make sure to turn off those settings.

What steps will reproduce the problem?
1.
2.
3.

What is the expected output? What do you see instead?
Should only send one keyboard message for the single down-up event.

Please use labels and text to provide additional information.

If object detection is successfully finished:
Use the Kinect depth map and OpenCV to detect the distance of object from 
Kinect camera.

Convert the PWM setpoints of the motors to speed, and calculate the integral of 
the speeds to track position.

Purpose:
Test messages from ROS cloud go to the right place.

This is an experiment and might complement or be incompatible with issue #43.

Explore the possibility of making the boundaries of the "is flat" zone 
dynamically adjustable during runtime.

This would not be automatically adjustable, just manually adjustable, but the 
idea is that the distance the robot prefers to keep from surrounding objects 
can be reduced when it is placed in confined areas.

For example, in a narrow hallway, the two diagonally vertical lines can be 
brought closer together to accept smaller clearances of either side.

Another example, is to be able to allow the robot to approach an opposite wall, 
such as in a T-intersection, closer than it would normally would be lowering 
the top horizontal line.

The video stream server currently displays a window with the capture image.  We 
want to add an option on the command-line to make it headless so that it should 
be able to run faster on the computer.

[deleted issue]

Integrate M-A's environment with the original 
GuardClient/HostServer/RobotClient code to be able to view the environment on 
the guard computer, and not on the robot itself.

Purpose:
Automated navigation of area

Robot needs to have a direct left and right sensor, to improve explorer 
efficiency

Purpose:
Implement OpenCV with Kinects as a ROS Node for visual interfacing.

Purpose:
Create and Test methods for secure connection between client and server.

If you close the MV server window, the program keeps running.

To work around the bug in Linux for keyboard events, we'll also use the Xbox 
controller.  This will also allow us to have smoother controller of the robot.

Robot should be able to move around on its to explore
- Moves forward
- Turns left or right when available
- Turns around when dead end
- Moves towards unexplored area

Purpose:
Automation

Explore the possibility of diving the main "is flat" zone into separate zones.

For example, the top portions of the diagonal edges can be replace with wedges 
(wide side at the top) to be able to detect mild drift of the robot.

We might want to divide the trapezoid with a horizontal line about half the 
high of the current area so that we can allow the robot to detect the opposite 
wall of a T intersection from a distance, slowly approach it, and turn once it 
is close enough.  Otherwise, the robot would refuse to get into the 
intersection because it would be afraid of getting to close to the T wall.

Change hardware controller to be able to communicate with ROS interface

Install the libfreenect software in order to communicate with the Xbox Kinect 
Camera and identify all available features

Replace the simulated rangefinders with the readings of the real ones.

The MV client-server is really slow (5 reqs/s max).  Make that faster.

Need to implement a simple proof-of-concept program for the ultrasonic 
rangerfinders once they are installed.

addition of multiple robot agents on sim, for future work

Set up kinect to analyse images

increase matrix size for Kinect flat floor detection.

Download openCV and install for Microsoft Visual Studio 2010

Develop an application that captures video using the built-in webcam of the 
laptop/computer

Create init scripts to allow RobotClient program and possibly stream_server.c 
program to start automatically when the computer in the vehicle boots up.

Test for task management

[deleted issue]

Explorer car still sideswipes walls and stop
-

slider to dynamically change the precision of the sensors

The server currently only accepts one client, and must be restarted if the 
connection dies.  Make the server able to receive multiple concurrent client 
connections, and multiple sequential connections.

Purpose:
Implement a Listener in ROSJava

The listener will be able to receive message from implemented robotclient 
(implemented in JAVA)

Implement a static map for demo to ensure a good working map

Purpose:
Implement a ROS simulators to develop and test without the use of hardware

Unit test currently slow down our progress (and we are in the final phase of 
the project), but if we ever continue this project they might be useful.

Angles of probe causes crashes

Currently, the machine vision needs to be interrupted, some code needs to be 
commented-out, and restarted to capture a new calibration frame.  We want to be 
able to do this while the program is running, on request.

Generates Map during Simulation

simulator map, and car explore sensor discrepancies

Set up test cases for robot explorer
- need to set up specific map

As continuous integration seems important we decided to add it for a part of 
the project even if it can't be done for all the project.

Convert the C# flat floor detection program into a TCP server so that the Java 
control systems can connect to it and query whether the floor is flat or not.

Make the presentation needed for Capstone.

Use openCV to develop an application that detects any object using a camera