Hi all, I was wondering if I could get some advice on what steering mechanism to use with an autonomous car I'm making? The car tops out at 6-7 m/s and is about 2 KG in weight. Length 45 cm and breadth 27 cm. I'm not considering differential steering as an option as my car has to have 4 wheels. I preferably want to use Ackermann's principal
Also I had some doubts about power supply. I am currently using LiPo batteries of my old laptop battery connected in series to power up the motors and microcontrollers (Arduino and Raspberry Pi 3). Is it enough or do I need something else?
@YaddyVirus - You might be able to get by with fixed caster wheels and large servos to create an Ackermann-style steering system without the steering knuckle, steering gear box, tie rods, etc.
I would caution you against re-inventing the wheel, though. If you're still in the design phase for your vehicle, I would suggest you look at hobby RC products to see if you can modify your design to fit a popular scale.
For example, this 1/5 RC car is 60cm long. It has a complete steering assembly already, which I would assume means that spare parts or replacement parts are available for purchase.
This is what I was hinting at in an answer to one of your questions - look for parts already manufactured that you can buy off-the-shelf. You'll have a lot more success buying parts that have already been designed and manufactured than you will trying to make them yourself.
Ultimately, I can understand the desire to want to make a car "from scratch", but automobiles are so highly multi-disciplinary (RC automobiles even more so) that you should really stick to what you know for a first pass, then work on redesigning your vehicle once you have a functioning "skeleton"
So I, as more of a "mechatronics" guy, I might purchase the RF transceiver rather than try to design that circuit on my own, and would probably buy linkages, bearings, actuators, etc. (as I'm suggesting you do), and then design my own control board and the physical layout of all of those components.
Regarding your power supply, first I would suggest you check the terminal and charging voltages of the pack you created to be sure that you're not violating absolute ratings on any of your components. As long as you're not over- or under-powering your equipment you should be fine.
Whether or not it's "enough" depends on the runtime you want from the pack. For this, I would calculate an estimated power draw (circuits draw XX mW, motors combined draw YY W, etc.), then compare that to the energy capacity of your batteries in Wh. If you are drawing 30W and you have a 10Wh battery pack, then you can expect to run for about (1/3h) = 20 minutes. This would at least give you a ballpark of what to expect
@Chuck the thing is I'm bit outsourced for this project. Moreover in India its very hard to find parts like of that buggy you pointed, and online stores like those don't ship to India.
Thats the only reason I wanted to make these parts from scratch
That's going to be difficult, to say the least, unless you take some un-traditional approach like using servos to turn the wheels instead of a conventional steering system.
But I got my school to back me up financially so, that shouldn't be much of a problem. Anyways can you please elaborate on the castor wheel and servo approach?
Yeah I realized that soon after I started this build...
The torque that you'll need for the servo is not going to be very easy to calculate, so if you've got the funding I'd go for the biggest servo motors you can find.
The ones I linked above might be okay. It looks like they're maybe 2-3 times a smaller servo, but they could be worth it depending on how much testing you're willing to do.
For example, you could buy 3 or 4 pair of servo motors to find out which ones work best, but then you're stuck with all the ones that didn't work out.
The torque will depend on the surface you're driving on, wheel size, speed, traction, etc.
If you've got narrow tires in sand then they'll sink into the ground and become very hard to turn.
But if they're wide tires with lots of grip on pavement then they could also be hard to turn.
You'll still probably need some custom brackets to attach the servos to the casters, but that should be significantly easier to make than a custom steering knuckle and (especially) the little bearings you'd need for the wheels.
I don't think it'll be that hard. My robot is quite small and lightweight as compared to robots which generally participate in that event. Being a school student I also don't have access to as sophisticated parts and sensors as my opponents (Graduate, under-grad students) do. I have focused on keeping it as agile as possible. So according to me a regular metal gear servo should work.
The castor wheel and servo idea is definitely going to be easier than a steering knuckle or something...
One final thing before I leave and dive into my homework (which has been piling up lately), can you suggest what rpm do I use for my motors? I am planning to use these - robokits.co.in/motors/high-torque-dc-geared-motor-900rpm
Would this motor be able to run from a L293D motor driver?
The motor's rated speed will impact how fast your vehicle can go; you'll want to use the radius of your wheels and find the circumference of those as $2\pi r$.
Anyways, you can figure out how far the vehicle will go in one revolution, and the motors give you revolutions per minute, so you should be able to calculate distance per minute by multiplying the two
But whether or not you have the necessary torque is more complicated; you'll need a drive cycle analysis where you evaluate the worst-case slope, terrain, and payload to determine if you have the capability to do what you want. You're probably okay, but again, to be sure on that you'd need to do more in-depth analysis.
Anyways... So calculate how far the robot will move in 1 rev of the wheel I intend to use and then figure out whether the motors are good enough or not, check necessary torque as well, then if the motors are good enough, just get a better motor driver and I'm good to go, right??
Yeah, I think so. Like I said, though, the drive cycle analysis can be a little tricky making sure you get everything accounted for. If you have any spare motors laying around you should try setting up a little experiment where you predict the torque required and measure what the test vehicle actually delivers.
Just to make sure you're predicting correctly.
Anyways, if you would like any help with more open-ended design questions like this, feel free to ping me. Any time you type @Chuck here I get a notification. I might not (probably won't) respond immediately, but I will respond.
Yeah actually I was experimenting with a pair of motors that I had, I scavenged them out so I am not sure of their technical specs but they seem to work fine, along with the L293D motor driver... just a bit slow. And oh BTW the motor driver you saw on Adafruit was just an IC, what I have is a complete module, but I don't think that makes any difference. Here is a pic of it - forum.arduino.cc/index.php?topic=394191.0
Yeah, but typically a "module" is just a pre-packaged system that includes the particular IC along with whatever associated components are required for proper operation.
Anyways, module or IC, it certainly won't be enough for those motors!
Not sure if this is the right place to discuss this, but I was thinking there could be some improvements made on tags related to drones. E.g., "multirotor" instead of or as a synonym for "quadrotor", or even adding a "drone" tag, which doesn't seem to currently exist.