IAR 2016 Source

Authors:

s1311631 : Angus Pearson

s1346981 : Jevgenij Zubovskij

Install/Setup

Use of a virtualenv is suggested, which can be set up using virtualenv VENV. Virtualenv is then applied using > source ./VENV/bin/activate and deactivated using (VENV)> deactivate.

Code targets Python 2.7, but makes use of some future features.

Requirements can be installed using (VENV)> pip install -r ./requirements.txt. Pip probably should be upgraded as well, using (VENV)> pip install --upgrade pip.

Redis

Download Redis and follow it's instructions for compiling, testing and installing. A redis-server instance needs to be accessible for this code to run correctly, it does not have to be running on the same machine.

The server instance should be run with the given redis.conf file:

$ redis-server ./redis.conf

                _._
           _.-``__ ''-._
      _.-``    `.  `_.  ''-._           Redis 3.2.4 (00000000/0) 64 bit
  .-`` .-```.  ```\/    _.,_ ''-._ 
 (    '      ,       .-`  | `,    )     Running in standalone mode
 |`-._`-...-` __...-.``-._|'` _.-'|     Port: 6379
 |    `-._   `._    /     _.-'    |     PID: 10826
  `-._    `-._  `-./  _.-'    _.-' 
 |`-._`-._    `-.__.-'    _.-'_.-'|
 |    `-._`-._        _.-'_.-'    |           http://redis.io
  `-._    `-._`-.__.-'_.-'    _.-'
 |`-._`-._    `-.__.-'    _.-'_.-'|
 |    `-._`-._        _.-'_.-'    |
  `-._    `-._`-.__.-'_.-'    _.-'
      `-._    `-.__.-'    _.-'
          `-._        _.-'
              `-.__.-'

Bear in mind Redis is an in-memory data store, so on-disk saves won't occur that often; Manual saves may be neecded to force data retention.

Plotting

(VENV)$ python -i ./data.py
>>> ds.plot()         # Live plotting
>>> ds.static_plot()  # plot existing data
>>> ds._purge()       # Clear Redis
>>> ds.get()          # return all data

Further reference:

(VENV)$ python ./data.py --help

Methods in data.py also have documentation

Command-line shortcuts are

data.py --server=<REDIS SERVER>
data.py -s <REDIS SERVER>

data.py --delete
data.py -d

data.py --plot
data.py -p

data.py --rospipe
data.py -r

ROS Integration

Internally we're using Redis to publish and collect data. data.py has a method DataStore.rospipe() that pipes from Redis' PubSub into ROS topics. It can be run as a method call, or with the command line switch --rospipe or -r.

ROS.org has detailed docs on installation and dependencies. For the whole ROS pipeline to work, and instance of roscore must be running locally, and rospipe() will publish to the topics /statestreampose, /statestreamodom and /statestreamdist, as well as publishing a TR named khepera. Rviz should be able to detect these topics.

Utilities (utils.py)

Reference Frame Transform

utils.relative_to_fixed_frame_tf(fixed_x, fixed_y, relative_theta, relative_x, relative_y)

IR raw sensor reading to distance

utils.ir_to_dist(reading)

Occupancy Grid

A DataStore instance has it's own instance of a GridManager, called og.

Generating a map

The mapper is run asynchronously from other programs, like so:

$ python mapping.py --map

Main Methods and Members

NB:

Full class names are given, though these methods and members should be accessed though an instance, like so:

>>> ds = DataStore()
>>> ds.push(...)
>>> ds.og.get(...) # 'og' is a DataStore's GridManager instance 

DocStrings should be consulted for fuller documentation

GridManager.get(x, y) : Gives occupancy - unoccupied to occupied [0..100], -1 unknown

GridManager.mget(l) : Gives list of occupancies

GridManager.granularity : Stores the granular unit size for the Occupancy grid.

GridManager._snap(coord) : takes an arbitrarily accurate number and returns the coord of the grid cell it falls within. You probably shouldn't use this (hence the underscore), as .get() automatically calls this on it's args.

Deleting data & cleaning up

The occupancy grid can be destroyed (keeping all other data such as poses) using the _destroy() method, e.g. ds.og._destroy(). This is different from the DataStore._destroy() method that will flush the entire database.

The destroy method can be called directly from the command line:

(VENV) $ python mapping.py --destroy

All listeners and programs using an occupancy grid should be restarted after a destroy.