Mitch Torri, the owner of ASI in Mendon, Utah, generously provided my V2R_AG_Robotics 4H club with a robot project reminiscent of the Chaos Robot they developed for the military in the early 2000s. Our current robot setup includes two electric wheelchair motors, a Sabertooth 2x32 motor controller, and is operated via a FlySky remote control. We aim to expand this system to accommodate a total of eight motors of similar size.
My primary question is: How can we configure multiple Sabertooth 2x32 controllers to work harmoniously within Synthiam ARC to control and collect data from all the motors for autonomous operation? It seems impractical for a single person to manually coordinate eight motors, especially across rugged terrain. We have a challenging task ahead, as we need to manage a 30-acre steep, overgrown hillside along the downtown Ogden River in Utah. The goal is to clear and maintain this area to prevent potential fire hazards, similar to those seen in Los Angeles.
We envision deploying a swarm of these Chaos-style robots, which will communicate with overhead drones. These robots will cut through three-foot tall grass, brush, and maneuver around both standing and fallen trees.
We plan to use gas engines to initially charge the batteries, but our future goal involves developing a more efficient power supply system. With various technological advancements coming together, this project is an exciting opportunity to embrace what I refer to as 21st-century literacy.
What are your thoughts on this endeavor? Is it a bold vision or overly ambitious?
Can you be more specific about "visual autonomous guidance hardware"? Are you asking about object detection, navigation or both?
Object detection: knowing there's an object in the direct path. (i.e., stop the robot while you're controlling it to prevent the robot from hitting something)
Navigation: involves knowing the robot's position in a 3D/2D space, knowing objects within that space around the robot, and being able to navigate around obstacles to a goal location.
I highly suggest you read the steps of the getting started guide. If you follow the first few steps of the getting started guide, you should be able to get up and running quickly. Here's the link to the first DIY step: https://synthiam.com/Support/Get-Started/how-to-make-a-robot/make-a-diy-robot
Essentially, you'd be doing this...
Add Sabertooth Movement Panel robot skill
Add a robot skill to make the robot move, such as a joystick, joypad, or something
Optionally...
Add a camera to the robot
Add the camera robot skill. Now you can see what the robot sees
Optionally more...
Use the Configuration menu of the camera robot skill to enable Movement Tracking (I'd enable the variable speed option as well)
enable a tracking type and have the robot automatically track an object
As for hardware, that's up to your requirements. The getting started guide has a section on choosing your EZB hardware.
I was reading through the comments here on my Questions and camera came up. I followed the link in the comment and it took me to real sense T256. So yes I need to get into the guide but I needed to know what hardware to get on order to move forward. You say the guide will help with that. I also noticed a comment about lidar. So while I had a minute just now my brain was seeking what is the current hardware to procure? I like all the points made, just as I follow the embedded links it takes me to obsolete hardware some of the time and I'm trying to get my head wrapped around the easy-b hardware. Like you say I need to get into the guide!
Intel did a limited manufacturing quantity of their RealSense product line, but they're handy if you get your hands on one. Just because something isn't manufactured anymore doesn't make it obsolete; it makes it more difficult to obtain. You will find that most hardware manufacturers of niche products do short runs because the market isn't large enough to maintain a manufacturing schedule. Intel may not manufacture Intel RealSense products, but Intel continues to maintain software development. I'm unsure what else you would find that's "obsolete", let me know; it might fall under the same category as the RealSense.
As you mentioned, there is no direct hardware replacement for the Intel T265's feature set. To achieve the same outcome as a T265, you must combine a few sensors and robot skills, such as a Lidar and The Better Navigator. But again, that depends on your environment, and I still don't know what your robot's goal is. It'll be more apparent to provide guidance when I see your robot's primary task.
As for hardware to purchase, that's up to your goal and outcome. The first step of the guide is to dump everything out of your head into a project plan. If you don't do that, your robot can be a playful testbed for experimenting, but it won't do a particular "thing". For example, if you set out to make a peanut butter sandwich, that's your goal. However, if you go into the kitchen with the idea of "just putting food in your mouth", who knows what you end up eating
So, having a plan is a good place to start for any project. You can always start with the "Goal" and work backward. Identify what hardware you'd need to achieve it. Then, identify what micro-goals you can split the project into.
For example, if you want the robot to navigate using sensors, you'll first need to identify what the robot is.
The guide is essential because you can work on each step. I'd set your first goal to "make the robot move".
Having it remote-controlled by a toy RC controller doesn't make the wheelchair base into a robot.
Having the wheelchair base move with instructions from a computer, even if controlled by a human, makes it an entry-level robot. This is because a computer will use algorithms and conditions for customizing the movement, even if controlled by a human. Once you have the robot controlled by a human through the laptop and EZB combination, you can add sensors to have the robot perform actions based on real-world environmental conditions.
https://youtu.be/aAmD7eqdF98?si=9XlgkvK91v8Kxnm_
This link is the target we are shooting for. The Chaos style robot however, with anarticulating system instead of just the tank dual track.