↑top

Connecting a Workstation and Starting the Robot

The Magni robot is normally running all the software for whatever sensors and navigation mode you are using for your project but there is no practical way to see what the robot is doing within ROS because the Host Linux Cpu running the Magni itself is headless and normally only has WiFi to access.

The workstation is setup so the robot is the ROS master and this allows ROS running on the workstation to monitor ROS topics and inspect the state of ROS using command line tools or graphical tools.

Your workstation may be a Ubuntu Linux system of your own such as a laptop, or you can use our preconfigured virtual machine in the form of a VirtualBox.

This page discusses connection of your laptop or workstation or the usage of a VirtualBox to control the robot. If you do not have a linux computer available you can still use a virtual Linux machine to connect to the Magni robot.

Additional information may be found here.

Reasons A Workstation Comes In Handy

This little section is optional reading to better understand the power of running a workstation.

The robot itself is normally running all hardware drivers for sensors such as any sort of camera or sonar sensors or even Lidar unit. If your system uses any form of navigation then things like fiducial navigation, gmapping, move_base, move_basic or AMCL all run on the robot itself so they are close to the hardware.

The prime reason a workstation or virtual machine is required in a system using a Magni is that there so a way to get access through console screens running as SSH terminals seen on the workstation to start, stop and monitor output from software running on the robot. A second fairly important console screen sort of use for a workstation is that since it will also see all the ROS activity the workstation can monitor ROS topics or other status for ROS at that time.

A workstation is also used so that graphical tools such as RViz or plotting software can be run on the screen of the workstation. This is the case if the robot is to do navigation you would then run visual tools such as the RViz environment to see the robot move around or to set goals. In this mode you could configure RViz to show you the sensors such as our sonar board or even a Lidar unit.

Another need for a workstation would be so that you can monitor values in a ROS topic either from command line tools or such as for when you want to do PID tuning it is important to see the PID error values in a graph you would tend to run rqt_plot on the workstation or just run image_view to see what the raspicam camera is seeing as the robot drives around.

We are not in this section discussing purely virtual simulation tools but those too can be run on a workstation assuming the workstation has a powerful enough configuration.

Use of your own Linux machine with WiFi

If you have your own Linux laptop already configured with ROS to be running the same ROS release as the robot then you can use it to connect to the hotspot on the Magni computer using standard Linux interfaces to connect to a network.

Skip ahead to the next major section 'Connecting to the Robot and Logging In' for laptop if you are not going to run a VirtualBox instance.

Using our out-of-the-box virtual machine workstation

You can setup a Linux VirtualBox to communicate with the Magni robot. In this case we have a Linux image that is preloaded with a great many ROS tools so this is a nice way to get going fast if you have the resources to be able to run VirtualBox environment.

In a later section we will explain how to attach the robot to an existing WiFi network and how ROS can take advantage of that to control the robot.

Connecting a Virtual Machine

If you are running under VirtualBox, you will have installed this virtual machine with a bridged network. Thus, the VM will see whatever network your host system is connected to. If your workstation is running, shut it down. Then connect your host system to the ubiquityrobotXXXX network. Now start the workstation (that is, the Ubuntu system running under VBox); it will be connected to the robot’s network ubiquityrobotXXXX. The password (sometimes called the security key) to connect is robotseverywhere.

Connecting to the Robots Network and Logging In

At this point you should have either your own linux physical machine or the VirtualBox setup described above and we are ready to connect to the Magni WiFi hotspot or sometimes for development the robot may be wired to your LAN.

Connecting To The Magni Robot WiFi hotspot

If you have received a Magni with the Raspberry Pi already installed, or loaded the default Raspberry Pi 3 image from downloads.ubiquityrobotics.com, the robot will boot up in WiFi Access Point mode. This is a WiFi mode which provides its own network to which you can connect your workstation. The SSID (network name) is ubiquityrobotXXXX where XXXX is a number letter combination. You should find it under network list; the password (sometimes called the security key) to connect is robotseverywhere. If you can’t find it under available networks, try restarting the robot.

Open A SSH Console Once You Are On The Robot’s Network

Once you are on the robot’s network, via WiFi Or Wired, you can connect to the robot itself. On your workstation, start a terminal window (Linux shortcut: ctrl-alt-t). In that window, log in by typing

ssh ubuntu@ubiquityrobot.local

The password initially is set to “ubuntu”.

If you have trouble using ubiquityrobot.local, use the IP address, 10.42.0.1 instead. In some environments, this works better.

You may see:

The authenticity of host ‘10.42.0.1 (10.42.0.1)’ can’t be established.
ECDSA key fingerprint is SHA256:sDDeGZzL8FPY3kMmvhwjPC9wH+mGsAxJL/dNXpoYnsc.
Are you sure you want to continue connecting (yes/no)?

Answer yes

Failed to add the host to the list of known hosts (/somepath/.ssh/known_hosts).

Ignore that. You will be asked for the password, which initially is “ubuntu”.

ubuntu@10.42.0.1's password:ubuntu

Now you are connected and logged in.

Welcome to Ubuntu 16.04.3 LTS (GNU/Linux 4.4.38-v7+ armv7l)

Running ROS apps not requiring a Display

Many of the robot applications we supply as demos or some you may be writting yourself do not require a laptop if they do not need to display anything. These types of command line applications can be run on a console to the robot.

A common simple app to control robot movement for example:

rosrun teleop_twist_keyboard teleop_twist_keyboard.py

Even some complex applications such as to start the robot’s navigation software can be run on the robot directly in a text console.

roslaunch magni_demos simple_navigation.launch

We would suggest that to run simple_navigation you should see our Fiducial-Based Localization Page as our intent here is not to explain that complex mode, simply to show an example of a ROS app.

Setup Laptop ROS Environment For Magni Control

In order to have your laptop be able to issue ROS messages to control the Magni such as running twist on the laptop or to run programs that require a graphical display such as rviz you will need to run ROS on your laptop and setup the laptop instance of ROS so that the robot is the ROS master.

Once the laptop (or VM) is configured with the robot as the ROS master you can inspect the ROS topics on the laptop and even view what the camera is seeing from your laptop or do other things like configure dynamic ROS parameters which all are done with a laptop display normally.

Rather than duplicate the required setup here you should refer to the ROS Workstation Setup Page to do these required configuration steps.

At this point you will be able to control the robot from the workstation keyboard or by using Robot Commander.

Doing a clean Shutdown Of the Magni Robot

Forgive us for first explaining how to shutdown and then power off the Magni robot but we felt this step may be missed if it were at the very end of this page.

We recommend doing clean shutdown on a Magni to lower your chances of corrupting your file system and to make your next powerup be faster by avoiding file system check next powerup. If you have the time to properly shutdown the linux host you will not run the risk, however small, of corrupting your Micro SD card.

Once you have an open SSH console to the Magni, which is explained next on this page, you can shutdown the Magni in a clean way using these commands and actions

If your Magni is a Magni silver and you have the sonar board you can always use the SW1 switch by holding it down for a couple seconds. This will lead to a clean shutdown of the host Raspberry Pi CPU.

If you do not have a sonar board you can wire for yourself a reset switch that is normally open and when connected connects pin 31 of P702 on the MCB to pin 30 of that large 50 pin connector.