EmbeddedRelated.com
Forums
The 2024 Embedded Online Conference

Regarding a specific microcontroller

Started by Raviteja_Banda 5 years ago12 replieslatest reply 5 years ago145 views
Hi,

I'm a masters student from Italy and I'm looking for a microcontroller with a some specifications, but i can't find one with those specs. My requirements are a microcontroller board with gyro sensors and camera interface with inbuilt debugger and proogrammer with RTOS installed. Can you suggest any board? Also my task is -

Take a picture with the camera and compare that image with already stored image catalogue in the microcontroller and also measure the angular velocity with gyro sensors. 
So, do I really need RTOS or can I also work without RTOS?
[ - ]
Reply by SpiderKennyApril 30, 2019

Why not just write an app for an iPhone / Android phone?

That will have gyro, accelerometer, camera, multi-core processor, debug tools, wifi, bluetooth and a whole lot more!

Alternatively, a Raspberry-PI with a "Sense hat" will come very close.

[ - ]
Reply by Raviteja_BandaApril 30, 2019

Hi, 

Can you suggest me any company which manufactures customised boards with my specs?

Also, I have gone through some STM microcontrollers and i found two different boards one with gyro sensors and other with camera. Can I use both of them together for the above mentioned requirements? Is it possible?

[ - ]
Reply by mr_banditApril 30, 2019

An RTOS is needed when timing is critical. From your description, this does not sound time critical. So - any "normal" OS would work.

If you don't want to roll your own, SpiderKenny's suggestion of a smart phone should work. I would go with the android because of the Linux OS.

[ - ]
Reply by Raviteja_BandaApril 30, 2019

Hi, 

Can you suggest me any company which manufactures customised boards with my specs?

Also, I have gone through some STM microcontrollers and i found two different boards one with gyro sensors and other with camera. Can I use both of them together for the above mentioned requirements? Is it possible?

[ - ]
Reply by mr_banditApril 30, 2019

You will need to do two things:

1. physically tie them together so that the gyro data is meaningful to the image data.

2. treat the gyro board as a slave to the image board && use some method of communication, such as RS232 (probably the simplest), SPI, I2C.

Please note I am assuming a fairly static test setup - take a picture, move, take a picture. Not something dynamic like an aircraft or an active vibration/rotate table. (If your ultimate goal is a dynamic situation, you should start with a static setup to work out all of the issues. A dynamic test jig will probably demand a custom data acquisition system.)

Can you give us some more details on your project and your goals?

[ - ]
Reply by rubenrayApril 30, 2019

choose a gyro sensor which already has the angular rate as an output see:. https://www.sparkfun.com/products/14001. its cheap and will leave your embeded systems free for the pictures related work. In such a case RTos is not needed. Of course always you can link any numbers of embedded system you want,  but the power  availiable marks the limits. (an imu sensor like the 14001 above includes a embebed controller)

[ - ]
Reply by mr_banditApril 30, 2019

Now *that* is a cool little board! I will need to remember that...

I wonder what the drift specs are... Remember: there is one constant: gravity sucks. I did a cheap 3dof tracker. That bit of physics made it possible.

I use boards like this to determine if a complex datapath is working. I have a video SERDES with a back-channel I2C bus. I put a simple I2C device (temperature sensor) at the far end to make sure the entire path works.

[ - ]
Reply by rubenrayApril 30, 2019

Drif specs refers to the error introduced by the integral (math) operations. anyway , the above sensor  includes a complex algorith called kalman filter which  corrects the drift .

[ - ]
Reply by JackCrensApril 30, 2019

I'm more than a little surprised that no one has given you the obvious answer: 

GET AN ARDUINO UNO.

They're cheap ... under $12 on Amazon, and under $5 for Arduino compatibles.  The OS is built in, and the IDE is free. You program it in C++ ... no assembly required.

The Arduino has neither gyros nor a camera, but it can accommodate just about ANY device. For quick and dirty, just connect the Arduino and the devices to a breadboard.  Otherwise, you can get what Arduino calls "shields" that plug into the Arduino pin headers.

The nicest part of using an Arduino is that the available software library is ENORMOUS.  Almost any device you can think of is not only Arduino-friendly, but almost certainly has a downloadable library already written.  





[ - ]
Reply by jkvasanApril 30, 2019

Hi


You can buy Synergy S124 (Renesas MCU )evaluation board that has accelerometers, etc and also has a perfect IDE which is integrated with Threadx OS. Beauty is, you get the RTOS free. I guess this could be your solution.

[ - ]
Reply by Raviteja_BandaApril 30, 2019

Hi all,

Thanks for all your replies. I appreciate all the help.

But the problem is I'm from mechanical and aerospace background. So, I have no knowledge in embedded systems. This is the first time I'm working on embedded systems. So, half the things you guys are saying are above my knowledge. Can someone suggest a simple solution in layman's terms. 

Thanks in advance.

[ - ]
Reply by mr_banditApril 30, 2019

Well: "Can someone suggest a simple solution in layman's terms."

You seem to want to successfully leap half-way across a canyon. 

What you want to do is somewhat sophisticated. It seems simple on the surface, but there be sharks in the water. Lots of subtle things to grab you by the ankles and drag you down. Honestly, what we have said above is pretty simple. I am sure you are good at what you do, but this is not a beginners project, especially with an RTOS. This is a Jr/Sr CS level project.

My basic suggestion is find a CS grad student. They should be able to do this fairly easily.

You also need to look at how you are going to fuse your sensor data together. Also not trivial. Not to mention how you are going to compare two images - that can be a graduate-level problem right there. The "Mark One Eyeball" is good for comparing pictures.

The board suggested by jkvasan might be a good way to go. (ThreadX is one of my favorite RTOS's.) Or a Rpi with a camera and the sparkfun IMU board.

Personally, this looks like a team project. It's the kind of think I would love to work on, but you cannot afford our rates

I hate to "harsh your mellow". I suggest you step back and think about what your real problem is && what resources you need to solve it. At a minimum, you could take the images and orientation data && dump them to an external drive, then post-process the data. This gives you a gentler environment (eg a PC) to do the data crunching. (Even if this was for a real product, that is what I would do as one of the first design steps.)

HOpe this helps. Success to you!

The 2024 Embedded Online Conference