Thanks to the Pioneer’s background as a research ‘bot, it comes equipped with a rich C++ API allowing for easy programmatic access to the microcontroller, via a serial link from the on-board PC. This has given us a great head start getting control software for equipment like the robot arm and motor control working quickly and effectively.
As mentioned in one of the previous posts, we’re using a simple 802.11-based wireless ethernet link between the lander site, the navbot and our Pioneer rover. Our ground control station will consist of a setup of laptops and large screens, allowing us to view the various video streams and telemetry from the rover, as well as controlling its movement, track deployment and arm control. Mobile Robots provide a small suit of control applications designed for use over a wireless link, which we have applied and combined with additional applications to allow us complete control of the rover while it is performing its mission (or at least thats the theory!).
We have tested these application thoroughly and had great success in remotely controlling the rover once out of visual range. So control in the crater shouldn’t be a problem.
So the Pioneer, while a great base for software development for rover control and autonomy, wasn’t really designed to undertake quite such a heavy set of physical modifications as those we have applied. As we’ve spoken of before, the large, wet-cell batteries were replaced with a smaller, low-profile Li-Ion battery in order to make room for the additional equipment we have designed. This was just the beginning…
Without going into specifics, the new rover has a refined drive system, track pitch drivers and a reshuffle of the other odds and ends the rover needs to run (the microcontroller and OBC…!). Let’s start with the base model.
The Pioneer 3-AT is a four-wheel drive, wheeled rover. Internally there are 4 electric motors, with two each driving a pair of wheels each via a belt. The remainder of the internals are fairly inefficiently stacked around these motors and the wet-cell batteries.
So, while our new battery saved some space, we had to fit additional motors inside to adjust the the tracks and still retain enough space to fit all the original equipment back inside! Here’s the mounting bracket with the motors attached, ready to be fitted inside the chassis.
Once securely in place we attach our replacement drive chains, providing a more compact and stronger drive system for the tracks. The remainder of the internals just about fit over these.
Our tracks are currently being tweaked to get the best performance and we hope to have them finalised early next week. The axles, while similar and in the same positions as the originals, have again been modified to allow the track structure to rotate independently of the track belts, not the easiest of problems to overcome, but so far our solution seems to be doing the trick!
While we’re furiously working away to complete the rover construction in time for the challenge, the majority of our other systems are complete. The next one we’ll talk about is the navigation system.
While this is clearly crucial to completing the mission the on-board vision system is relatively basic. We are using a stereo-camera to provide a live video stream back to our ground-station/control centre. This will be our primary vision system and used to navigate the rover, search for the samples and aid the robot arm when collecting them. In order to make full use of the capabilities this provides we have mounted the camera to a pan and tilt unit. Coupled with a pair of bright headlights we will be able to use this unit to pan around the crater floor once inside to search for the samples, without wasting additional power driving around the site.
The video below shows the PTU with the camera and headlights attached, in both lit and and unlit conditions.
And here’s our dev-bot with the arm attached on the back and the camera on the front. These will be next to each other on the final rover however this allowed us to check the camera was capable of giving a clear view of the arm workspace. It looked great!
The devbot rover with arm and PTU/camera/headlights attached. The combined unit did a great job of illuminating and videoing our lab!
Also visible in the photo above is the prototype part for the secondary area of our navigation system, the long bar with coloured LEDs at each end. These lights will be attached at the very front and back of Selene, and will allow us to use our relaybot camera to orientate ourselves back toward our entry point once we have collected the sample.
The relaybot, with the camera visible mounted on the front
The principle is fairly simple, while the primary purpose of the relaybot is to allow for wireless communications between the lander site and Selene once inside the crater, the fact that it will have some element of mobility will allow us to position the relaybot on the edge of the crater rim, where we estimate that it’s field of view should cover the entire site we have to explore. Once we have located and collect the soil sample the relaybot camera will detect the alignment of the red and green LEDs mounted on Selene. Once directly pointing at the relaybot we will have a clear indicator that we are pointing at our entry point, and therefore the optimal location to exit (in order to achieve maximum points!). We anticipate becoming easily disorientated and lost after searching the crater floor, this system will allow us to quickly regain our bearings and direct the rover back home to the lander!
The Pioneer 3 is powered by 1 to 3 lead-acid wet cell batteries. These provide a nominal 21AHr of energy, and under normal use should provide power for Selene well over the 2 hour time limit for the challenge.
However, they are particularly large and heavy. So thanks to the guys at Lincad, we now have a Li-Ion battery pack roughly half the volume (casing aside) and less than half the weight while retaining the same 21AHr capacity of the originals!
The new, smaller Li-Ion battery underneath the original Lead-Acid cells.
The new battery provides a number of benefits, the most significant of which is the more convenient shape. The new battery sits in the belly of our chassis, lowering the center of gravity and providing easy access through the bottom of the rover. The smaller size leaves room for the additional flipper control mechanisms inside the chassis.
It’s been quiet on the blog recently as we’ve been hard at work getting our rover up and running ready for the challenge on the 20th October!
We’ll be posting some photos later but here’s a quick run down of our progress to date…
The track lifting mechanism is now completed and working, using our microcontrollers we can stow and deploy the tracks as required depending on the terrain. Due to the total size of our tracks we just now have to wait for the (patient, brilliant) workshop guys to drill and tap the hundreds of holes needed to attach our grousers to the belts!
The navigation system has been completed, and works in two parts. We have a stereo camera mounted on a pan-tilt unit for our primary navigation once on-site. We’re ironing out a few bugs but its nearly there. The second part is a “home-beacon” system, incorporating our secondary rover/relay station, positioned on the rim of the crater. As well as acting as a communications relay the rovers on-board camera will detect coloured LEDs on the primary rover in the crater, and the ground-station software will use this to help orientate the rover toward our entry/exit point on the rim. With the dark environment this is likely to be a crucial aid in the task of returning the sample back to the lander site.
Our communications system is also now completed. Based on 802.11b/g wireless ethernet, we have successfully completed remote teleoperation and range testing of our development rover, with a total record distance of over 300m, non-line of sight connectivity from the control site! We also incorporated endurance tests of our new, smaller and light-weight Li-Ion battery pack, and were thoroughly impressed by the results.
More info and pictures will be coming up, and we’ll be posting more regular updates as we approach the challenge date.
We’re working hard now to finialise our track deployment system. Unfortunately the Mobile Robots Pioneer 3-AT we’re using isn’t too accustomed to being rejigged inside, so we now have a large Pioneer jigsaw puzzle!
Pioneer Jigsaw Puzzle
This required some “delicate” modifications to achieve!
Gently does it...
Now we have individual components we can look at how to rearrange them inside the Pioneer chassis. We’ve done plenty of modeling using CAD to give us some confidence that our modifications *can* fit inside, however it is extremely difficult to tell without trying it on the actual rover, those cables really take up a lot of space! We’re looking at replacing our big, heavy lead acid batteries with much smaller Li-Poly ones, but other than that we need to rearrange all of the existing rover equipment to a way that allows us to fit the track pitch control motors and gearing, a slightly difficult feat due to the requirement for an axle to pass across the entire width of the rover at both the front and back.
We’ve also started range testing the 802.11-based wireless equipment we plan on using to control Selene once its out in the field. This has started with the nanorover we want to use as a comms relay/video link. With some new, high-gain antennae attached to my laptop and the Surveyor SRV1 we managed to boost the link strength a significant distance over its poorly performing standard configuration. It’s difficult to see in the picture below but the tiny rover made it all the way to the bus and was still going strong, giving us a live video link.
Long distance shot
Attaching a similar antenna to the Pioneer gave us a range of nearly 200m line of sight! Check out the pictures below.
A large chunk of our campus for reference
The SRV1 nanorovers 70m range
The 180m the Pioneer got before we ran out of field!
That’s it for now. Stay tuned for more updates soon!