Use the Intel Realsense T265 tracking camera for way-point robot navigation
+ How To Add This Control To Your Project (Click to Expand)
Manual
With its small form factor and low power consumption, the Intel RealSense Tracking Camera T265 has been designed to give you the tracking performance for your robot. This ARC user-friendly robot skill provides an easy way to use the T265 for way-point navigation.
The T265 combined with this robot skill provides your robot a SLAM, or Simultaneous Localization and Mapping solution. It allows your robot to construct a map of an unknown environment while simultaneously keeping track of its own location within that environment. Before the days of GPS, sailors would navigate by the stars, using their movements and positions to successfully find their way across oceans. VSLAM uses a combination of cameras and Inertial Measurement Units (IMU) to navigate in a similar way, using visual features in the environment to track its way around unknown spaces with accuracy. All of these complicated features are taken care of for you in this ARC robot skill.
Update Firmware
The device sensor may require a firmware update.
1) Visit the Realsense GitHub page, scroll to the bottom of the page, and install the Intel.Realsense.Viewer.exe from here: https://github.com/IntelRealSense/librealsense/releases/latest
2) Click the hamburger settings icon and select Install Recommended Firmware
Robot Skill Window
The skill has a very minimal interface because it pushes data in the NMS and is generally used by other robot skills (such as The Navigator). 
1) Drop-down to select Realsense device by the serial number. This is useful if there are multiple devices on one PC.
2) START/STOP the Intel T265 connection.
3) The confidence of the tracking status between 0 (low) and 3 (highest). In a brightly lit room with many points of interest (not just white walls), the tracking status will be high. Tracking will be low if the room does not have enough light and/or detail for the sensor to track.
4) Log text display for errors and statuses.
Config Menu
1) Mounted Radius Offset (mm) is the distance in mm of the T265 from the center of the robot. A negative number is toward the front of the robot, and a positive number is toward the rear. The sensor must be facing 0 degrees toward the front of the robot. The sensor must not be offset to the left or right of the robot.
2) Enable Video Stream will send the fisheye b&w video from the T265 to the selected camera device. The selected camera device robot skill must have Custom specified as the input device. Also, the camera device will need to be started to view the video.
3) Distortion Correction will use a real-time algorithm to correct the fisheye lens, which isn't always needed and is very CPU intensive.
Video Demonstration
Here's a video of the Intel RealSense T265 feeding The Navigator skill for way-point navigation
ARC Navigation Messaging System
This skill is part of the ARC navigation messaging system. It is encouraged to read more about the messaging system to understand available skills HERE. This skill is in level #3 group #2 in the diagram below. This skill contributes telemetry positioning to the cartesian positioning channel of the NMS. Combining this skill with Level #3 Group #1 skills for obstacle avoidance. And for Level #1, The Navigator works well.
Environments
This T265 will work both indoors and outdoors. However, bright direct light (sunlight) and darkness will affect performance. Much like how our eyes see, the camera will is also susceptible to glare and lack of resolution in the dark. Because the camera visual data is combined with the IMU, the camera must have reliable visible light. Without the camera being able to detect the environment, the algorithm will be biased to use the IMU and will experience drift, which greatly affects the performance of the sensor's accuracy.
Screenshot
Here is a screenshot of this skill combined with The Navigator in ARC while navigating through a room between two way points.
Starting Location
The T265 does not include a GPS/Compass or any ability to recognize where it is when initialized. This means your robot will have to initialize from a known location and direction to reuse saved maps. Make sure you mark the spot on the ground with masking tape where the robot starts from.
How To Use This
1) Connect your Intel RealSense T265 camera to the computers USB port
2) Load ARC (version must be >= 2020.12.05.00)
3) Add this skill to your project
4) Now we'll need a Navigation skill. Add The Navigator to your project
5) Press START on the Intel RealSense skill and data will begin mapping your robot's position
How Does It Work?
Well, magic! Actually, the camera is quite interesting and it breaks the world down into a point cloud of features. It remembers the visual features so it can re-align itself on the internal map. It uses a VPU, is what Intel calls it. Here's a video of what the camera sees.
Related Content
A Little Of This, A Little Of That
This seems like a really primitive way to do this but environments change (pallet moved in a warehouse etc) so you need something guaranteed to lock on to so you can move along the next edge to the next way point. We need a better solution than this in the future to deal with drift, the need to recalibrate and the need to enter into a known path. I guess ML is the long term answer but in the short term the world will be filled with fiducial's.
About 11 min mark onwards.
Put it this way, the problem of a robot knowing where it is in the world is so big that Boston dynamics merely threw a bunch of redundant vslam LOL. And most likely a bias shifting complimentary filter fusing the sensor data. The sensor that starts to drift first would begin getting a lower weight in the fusing calculation.
Anyway - the much larger conversation on this topic is not about sensors, but AGI. Artificial general intelligence.
Even if you close your eyes and I move you to a new location, you’ll have no idea where you are. The robot needs some level of self in relation to the world to create an adaptive internal model. Like how we do
Question is there a way to tell robot it is currently in a certain location when it drifts. I start in the docking station and turn Realsense on. Then I tell roomba to move back about 1 foot from docking station and then give waypoint voice commands. It knows where it is because I started realsense when it was docked. Now it runs around my house but starts to drift. I guess I could send it back to a point in front of docking station and then dock using ControlCommand Seek Dock on roomba then stop realsense, then start again then backup one foot again etc, but it would be good if there was a way to tell realsense I am here. please reset your tracking location to hear now?
We talked about glyph as being a way to fix with camera. Maybe if we had 2 cameras working we could triangulate and get position. I thought about some lines on the ground in various locations that the robot could follow and when he crosses them he starts to follow and finds a dash code etc to recalibrate using a sensor like this https://cnclablb.com/adjustable-line-tracking-sensor-module.html But I think what I need is a way to send a command to Navigator to say HELLO I AM HERE. Maybe something like a ControlCommand("The Navigator", "ResetWayPoint", "Kitchen");
You see, the T265 requires a cartesian X/Y and Heading Degree for mapping. This means you will have to specify the absolute exact coordinate (including heading degree). I'm sure there's a way you can calculate it with a glyph size and such. But be forewarned that it won't be as simple as you may think.
*PS Note: moved the above comment into this thread because it's regarding the Intel T265 and now a robot build.
Question Can you Disable and Enable a USB port in ARC to reinitialize the driver or can you call a batch file to potentially use something like devcon to disable port in windows and then re-enable the USB port so it resets each time you stop and start it?
https://github.com/ralight/usb-reset/blob/master/usb-reset.c
example for my device
pnputil /disable-device "USB\VID_8087&PID_0B37\952322111114"
I could always just create a USB hardware add-on that allows me to turn it on and off with a relay (this is actually starting to sound like easiest method)
With regards to drift, I played with offset for a while and concluded -160 was a good level (When I go into a spin it doesn't create a donut) but yeah drift is still a problem.
Ryzen 5 must chew some power. What are you using for battery. I have blown out 4 Lipos running all this stuff (My 4S is now a 3S and my 3S's are now 2S's). I have now switched back to an old Sealed lead battery and plug in when docked so computer is always on. This seams to work OK and I get about 2 hours of play time. I am going to add some springs and magnets so it will auto charge (just a matter of finding parts) this week.
I really want to get this as autonomous as possible.
My setup with this mboard is recent and i am still in the build phase. I use 2 Xunzel SOLARX-8 gel batterys in parallel.
I need to work on distance sensors maybe laser sensors as these ultrasonics are garbage. Will smash into something or see something that is not there. Also I think we need a way to navigate with individual sensors, if something towards left turn right, towards right turn left. directly in front stop. When you have sensors directly on side you really don't care as much about distance from wall beside you, also when you are going fast they don't work as well as slow.
I also really need to add a panic button off an Arduino accelerometer to shut off roomba battery if it starts to go crazy. That seems to happen a lot :-)
Code:
Makes you wonder why the original Roomba smashes the furniture with all these sensors on board.
Thanks for the sample code and your idea to navigate around stuff using the onboard sensors @proteusy very cool. I had a play and if you run around really slow you can avoid some stuff but it still managed to knock over my green screen, dive under the treadmill, smash into my chair and get tangled up in the power cords under the desk. How a Tesla navigates at 100km an hour blows me away.
That said, I think I need to add a bunch more sensors, a LIDAR and lots of code to make sense of all this data, if wheels are spinning and it is not moving, I am probably tangled up have smashed into something.
Also, this robot skill will only display tracking cameras in the drop-down.
- Bug fix and small performance improvement