Autonomous Solar Panel Spider Project (Arduino)

Community General Discussions
1

Hello fellow robot enthusiasts/experts,

My name is Michele (Michael) and I'm from Italy. I really appreciate all your efforts and creativity for the love of robotics! It's truly amazing to see all your projects.

I am writing here today because I was inspired by all these robots and decided to make one myself. I am a Software Engineer so I don't really see any issue in programming itself. I have used and programmed LEGO Mindstorms in the past so I do have a bit of experience, but my true weakness is the Electronics itself. I really don't understand that much about Volts, Watts, switches and whatsoever. Shame on me, because while I was still in school I did use all kinds of stuff to practice on Assembly 8086.

Well, enough of that. I bought an Arduino Uno Kit and 2 Ultrasound sensors to start practicing. I also bought 2 DIY Spider Kits (Klann Walker) because this mechanism seemed to me the best and most efficient way to go through different terrains with only 2 servos.

What I'd like to do with it is:

- Expand the batteries from 2 AA to 4 AA.

- Make it wander and avoid obstacles autonomously.

- Collect data and/OR capture images and send it back to me via WiFi.

- Ultimately, I would like to power it via Solar Power, making it search for light when batteries are running low.

- Tell it to move to a specific GPS coordinate.

Now, my question is: Am I looking too far beyond this little robot's capabilities?

I really don't know how much power such tasks would drain and I don't know the solar panel size required. I am also probably ignoring a lot of considerations (for example, if I need resistences, if everything works with the same voltage, if there is a risk of burning something without certain components) that you people with more experience could point out, so if you have any thoughts and/or opinions that could help me feel free to comment and write. I'm pretty sure most of these things I'll figure out myself while I keep advancing through this project, but for now I just need to know everything I can.

Thanks,

Michele.

2

Welcome!

Hi Michele,

And welcome to this community! I like the way you want to start, small and then growing things up. The NiMH 4 AA batteries are a good improvement now, safe, cheap enough and re-usable.

From the beginning I think it’s better to have something “moving” (the robot), then adding sensors and some intelligence, later power management, and finally GPS…

I know it will take a while to develop all, but on this way you can also develop your robotic skills by little easy steps and consolidate your robot. On this way your new knowledge will be more solid, also your robot :wink: Good luck and I wait for your future updates.

3

Solar power is a interesting
Solar power is a interesting addition, but I would put it on much later, perhaps on another robot, when you understand electronics more.

Different battery chemistries require different types of charging circuits.

4

Any other considerations?

Would weight be a problem?

 

Also, How much would the batteries last considering 2-3 servos, a few sensors, a WiFi shield and/or a camera?

I don’t plan on rushing, but this is How I would like to advance while the project grows and so do I along with it.

5

About weight and related issues

Hi Michele,

If you use a lightweight camera it shouldn’t be a problem for weight, but it can be for power, even more if you add any kind of wireless transmission for that video signal (plus issues to power supply voltage to the transmitter as they may need 12V DC). Then the whole robot working time with 4 AA NiMH batteries could be something like 30 minutes, you can calculate like this:

2 servos: 400mA (600 mA peak)

1 “small” Arduino microcontroller board: up to 80 mA

1 little camera: 400 mA

2 small sensors: 40 mA

1 wireless video transmitter: 200mA at 12V! If power need is: 2.4 W then with a voltage booster at 0.7 efficiency at 6V (4xAA batteries), it will need: I = (2.4/6)/0.7 = 571 mA

1 camera: 50 mA approx

Total current consumption approx: 1541 mA

For a battery pack of 1000 mAh it means: 38 minutes working. You have to count on losses and not perfect batteries so time of working could be less…

 

 

 

 

6

Only my opinion While my opinion of Franciscodr’s estimate

is that he is low. Let us accept his numbers for the time being. You would have inefficiencies between the solar panel and the recharging circuit. Let us say for grins that you use NiMH, A) you will require 5 NiMH cells to get 6v 5 x 1.2v = 6 B) I believe you need to spec the solar cell to provide more voltage than the 6v the batteries should supply. Let us guess 7v @ 2A. If you look for a 14W panel you will see they are not tiny.

I may be completely wrong on my guesses, but, solar powered bots are few and far between due to efficiency issues. Solar panels are like 15% efficient. Charging circuits are maybe 80+% efficient. Buck/Boost converters are 80+% efficient. You have to over-build to make everything work that you need a larger machine than you were expecting.;

I believe you would be best suited to trying to make a battery powered system function first and gain some of the hardware understanding then transfer that knowledge to a different platform.

However, you mentioned moving across obstacles. One of our members has shared a 3d printable wheel that allows for transforming that might be a better fit to multi-terrain use. If you work with something like that and a moderately sized horizontally flat platform, you might be able to accomplish what you are after.

7

Spider
I also have this walker. I currently use 4 1.2V rechargeable AAA batteries with a pic24 controller. The motors are rated at 3V but run cool at 5V. My cheap sonar modules ($3 US) stop working if the voltage drops below 5V which occurs after a few minutes of running.

Planned modifications:

5 AAA batteries to get 6V,
5V motors,
NRF24L01 radio,
Encoders for motor speed feedback,
IR Distance sensor, and
Gyro/Accelerometer to detect motion.

Please post progress.

8

Walker motors
I just noticed that you say “Servos”. The spider motors are simple DC which require an h-bridge device and PWM to control.

9

You can use continuous
You can use continuous rotation servos, which are basically motors with the gearbox and controller build-in.

10

You plan sounds great, I
You plan sounds great, I think most of it is feasible, except for the solar power.

You see, running some simple LCD display, shining a LED or making sounds – the things we usually see solar panels used for – are relatively low power things. Anything involving mechanical movement is several orders of magnitude more power-hungry.

One advantage of using the Klann mechanism is that you don’t really have to carry the robot’s weight, it’s almost like using wheels. So you don’t have to worry so much about the weight of the battery and the solar panels. Another thing is that the robot can sit there for a long time without consuming any power, waiting to recharge. This allows you to have something like the Symet BEAM robots: https://www.youtube.com/watch?v=Ib3iUIr__0U – charge for 10 minutes, move for 10 seconds.

All things considered, I think the best way of making your robot “independent” and able to run on its own for long periods of time is to build a charging station, and program it to return for charging whenever its batteries go low. That’s what the commercial robots, like Roomba, do, and it’s the most robust system so far (except for the most obvious way, dragging a cable behind you).

Apart from power, navigation is actually a huge problem. Especially when you need to find a relatively small charging station and couple with it properly.

11

Would it be

Would it be able to power itself if I gave it let’s say, 1 minute of wandering and 10 minutes of recharge then? I’d like it to “patrol”, but not necessarily forever. It could just walk around only a few minutes every hour let’s say… 

I’d enable wifi camera transmission only when needed via IP? Power drain would be much reduces I suppose.

I also thought about the Roomba, but making it “slip” or connect to something is a very hard task, also because I need to tell it to return to same exact spot each time, and I’m afraid at this level it would be too inaccurate.

Thanks,

D

12

So…

First of all thank you very much for your help and ideas…

You’re right, sorry. I am totally ignorant on a few things but I am learning quickly. The Klann Walker consists of 2 DC motors and I would like to apply a servo as well for a 180 degree ultrasound sensor.

Would an L293D be enough to control the DC? For the PWM I just have to use some Arduino Motor controls libraries?

On the website it says Operating Voltage 1.5V and Operating Current 180mA… Does this mean it only draws 180mA/h? NiMH batteries are 1.5V circa so I should be fine right?

I also don’t understand how much amperes one of those AA batteries could provide. Around 2100mA, right? In that case wouldn’t a 4 battery pack be able to drive a basic OAR for several hours?

Pardon my noob questions, but as I said I really don’t understand the electronics logic. Am I calculating this correctly or is there something I am surely missing out?

Thanks in advance,

Michele.

13

Recharageable AA batteries

All my NIMH batteries are rated at 1.2V. Fully charged they seem to be 1.35V. Four of them can supply about 5V for a short time. Your processor board has a voltage regulator (probably linear) designed to reduce the battery voltage to around 5V. It requires some head room,

You can get boost regulators that will step up the battery voltage. With one of those you could use a single 3.7V LiPo or 4 AA’s. A simpler solution might involve switching to only 3.3V devices.

I recently purchased a couple of wireless induction coils from DFRobot to try non-contact battery charging. It was intended for another purpose but perhaps it wold fit the spider.

14

Could non-contact battery

Could non-contact battery charging be effective enough? Or is it inefficient?

I know smartphones get charged with that type of tecnology as well.

15

A raspberry model A with a

A raspberry model A with a wifi dongle may do the camera and driving work, with low energy costs. If you have a camera, a 5mW laser and not worry about time you can even do some ranging and localization using the camera and a computer vision library like openCV.

And welcome! I’m from Italy too, Torino:-)

16

I have no idea about how
I have no idea about how much walking to charging ration you would get. It’s very hard to tell without actually making some experiments, or at least having data on the solar cells that you are going to use.

WiFi transmission, while quite power-hungry compared to blinking a LED, is relatively low-power compared to the DC motors. Movement is going to be the most energy-consuming thing in there.

17

The Roomba iCreate has no
The Roomba iCreate has no problem going to its charger. It’s built in. Likewise if you star with any other Roomba, you can use the functions that make them go to their charging stations.

18

Raspberry Pi is a power hog.
Raspberry Pi is a power hog. I have one robot (https://hackaday.io/project/3301-pico-kubik-quadruped-robot) that uses one of those wifi router chips, running OpenWRT. It takes about 300-400mA at maximum, and is powerful enough to stream low quality video (although not enough to analyze it). A single 300mAh battery is enough for it to walk around (carrying its own weight) and stream video for about 20 minutes. Then it’s another 10 minutes to recharge.

19

A raspberry pi model A would

A raspberry pi model A would take about as much that, running headless and a wifi dongle. Anyway i’m looking for that chip:)

20

It is my understanding via another hackaday post

that the main reason RPis are so power hungry is due to their voltage regulator.