Kalman or not Kalman?


I haven’t posted updates in a while, we had one week of vacation to work on our project. Some things have already been explained in other post so here is a little summary.

We achieve to open the unbreakable shell of a sphero and look what’s inside. We then tried to guess which components are used for each functions: micro processor, accelerometer/gyrometer, magnetometer and so on.

The pcb is easily separated from the mechanical base (unless you unwillingly strip a screw and try everything in the world to get it out without destroying the board)

Be careful when you unscrew things..

Be careful when you unscrew things..








A good surprise was that the mpu used is exactly the same we worked with in class: the STM32F103CBT6, what a relief!
We want to reuse sphero’s base with our own pcb so we have to decide wich functions we want to use that sphero can do. Here is what’s we found on sphero that we might use:

  • The induction charger with 2 li ion cells, and the device to manage the charging.
  • The IMU (Inertial Measurement Unit) uses an accelerometer/gyrometer (3 axis each), we believe it’s the same that in the Galaxy s3!
  • A 3 axis magnetometer, apparently it’s there but we couldn’t identify the part. After investigation we found that orbotix decided not to use it because of the magnetic noise procuded by the motors  link here.
  • The bluetooth chip.
  • External cristal oscillator for the mpu.
  • The mpu: a STM32F103CBT6
A nice piece of work!

A nice piece of work!


Now our problem is to choose whether or not reuse the same parts for the same functionalities. Why bother to search for something else if the part they use does it’s job well?

Stay tuned to a further update on how we will choose the components.20130311_132352









2 comments to Kalman or not Kalman?

  • drix

    You might have said it in another post but how do you plan to improve the existing sphero ?

  • loikled

    We have some ideas but we wanted to better know what sphero is able to do and how. One of the main (and certainly the most difficult) objectives is to improve the controlling/navigation of the sphero. We found that it’s not that easy to remote control. But we have no idea of the complexity of the software in action to balance and control it. That’s why we would find sufficient to achieve the same level of control.
    Another aspect we want to explore is communication between several robots. Maybe make 2 modified spheros and make them do things together autonomously or have our rose rolls take control of several spheros without the need of a smartphone.

    We are open to anny suggestions 🙂