Prerequisite

Follow https://gitlab.com/EECE-5698-Group-7/vacuum-repo to root your vacuum (use imagebuilder.sh at your own risk).

Install python3-minimal using apt.

After rooted, need to start the vacuum in unprovisioned mode (by pressing the reset button on vacuum lightly).

Unprovisioned mode setup (if not included in rooted image)

1. in /etc/rc.local, add the following line

/opt/unprovisioned/start_wifi.sh >/dev/null 2>&1

2. create a foler at /opt/

mkdir /opt/unprovisioned/

3. create file /opt/unprovisioned/wpa_supplicant.conf:

network={
    ssid="<ssid name>"
    scan_ssid=1
    proto=RSN
    key_mgmt=WPA-PSK
    pairwise=CCMP
    group=CCMP
    psk="<password>"
}

4. create file /opt/unprovisioned/start_wifi.sh:

#!/bin/bash

file="/opt/unprovisioned/wpa_supplicant.conf"
if [ ! -f "$file" ]
then
    echo "$0: File '${file}' not found." > "/opt/unprovisioned/log.log"
else
    #add enough time to fix wrong wireless settings
    sleep 200

    #disable accesspoint
    ifdown wlan0 > /dev/null 2>&1
    ifconfig wlan0 down > /dev/null 2>&1
    killall hostapd >/dev/null 2>&1
    iw mon.wlan0 del >/dev/null 2>&1
    create_ap --stop wlan0 > /dev/null 2>&1
    killall wpa_supplicant >/dev/null 2>&1
    killall dhclient >/dev/null 2>&1

    ifconfig >> /opt/unprovisioned/log.log

    #login to your network
    ifconfig wlan0 up >/dev/null 2>&1
    /sbin/wpa_supplicant -s -B -P /var/run/wpa_supplicant_1.wlan0.pid -i wlan0 -D nl80211,wext -c /opt/unprovisioned/wpa_supplicant.conf -C /var/run/wpa_supplicant >/dev/null 2>&1
    echo "done!" >> /opt/unprovisioned/log.log
    dhclient wlan0 >/dev/null 2>&1
fi

5. Restart vacuum

Controller Tool Usage

python controller.py

Update scripts on vacuum

If in any case the script has been modified, we need to use init_vacuum.sh to update the scripts on vacuum. Run it within the python script via update so it takes in the discovered vacuum IP.

Otherwise, you can also set environment variable MIROBO_IP and run ./init_vacuum.sh directly.

Control vacuum

Type control to enter the control panel

Available commands can be found via help in control panel:

control >>> help
Control Command Menu
help                         - this message
home                         - move vacuum to dock location
status                       - print the status of vacuum
start                        - automatically start one cleaning sesssion and get data
move auto/pause/stop/home    - auto scanning movement (no data parsing)
move rotate speed time       - move (-180, 180)deg at (-0.3,0.3)m/s for `time`ms
fanspeed integer             - set fan speed to be [1-99]
goto x_coor y_coor           - move to x,y location on map
trace on/off                 - manually start/stop collecting trace
download trace/map           - download the trace or map on vacuum
config <cmds>                - configuration
quit/exit                    - exit controller (Ctrl + D does the same)

Example commands:

python controller.py
>>> control help
...
>>> control status
<VacuumStatus state=Cleaning, error=No error bat=93%, fan=1% cleaned 18.2775 m² in 0:16:12>
>>> control
control >>> help
...
control >>> start
Cleaning old data on device..
Enabling trace on the vacuum..
Starting..
<VacuumStatus state=Charging, error=No error bat=98%, fan=1% cleaned 2.6975 m² in 0:01:58>
...
control >>> config get remote_script_folder
/mnt/data/exp

Config vacuum

Nothing much yet. But notice once you entered config panel, have to exit before doing any controls.

config >>> help
Config Command Menu
help             - this message
set <key> <val>  - set key value in config
get <key>        - get config from key, if `key` not set, print all
save/load <file> - save/load configuration from file (default: ./config.json)
quit/exit        - exit controller (Ctrl + D does the same)

Example commands:

>>> config get remote_script_folder
/mnt/data/exp
>>> config
config >>> get remote_script_folder
/mnt/data/exp

Processor Tool Usage

After we get the .pcap file collected from phone, put the downloaded files (*_map.ppm and *loc.csv) with the .pcap file together under the same folder, e.g., example.

Then run

python preprocessor.py <folder path> --pickle -vd --orient 0

It will automatically extracts all RSS values from pcap file for each MAC address, and plot the .png figure with the RSS values being the red alpha channel (opacity) on the trace.

Commands details:

usage: preprocessor.py [-h] [--map] [--pickle] [--filters FILTERS]
                       [--sampling] [--sampling-num SAMPLING_NUM]
                       [--visualize] [--visualize-dump] [--orient ORIENTATION]
                       folder

Data Pre-Processor

positional arguments:
  folder                Specify folder path of data

optional arguments:
  -h, --help            show this help message and exit
  --map                 Enable to generate map images with scanned floorplan
  --pickle              Enable to dump into pickle images
  --filters FILTERS     Set filters to extract only 0: >, 1: v, 2: <, 3: ^, 4:
                        <^>, 5: <v>, 6: all
  --sampling            Enable subsampling to generate more data
  --sampling-num SAMPLING_NUM
                        If subsampling enabled, set the number of random
                        samples performed
  --visualize, -v       Enable to visualize map images while dumping to
                        pickles
  --visualize-dump, -vd
                        Enable to dump images while dumping to pickles
  --orient ORIENTATION  Specify orientation of the map

Reference

1. dustcloud
2. aerodust