tariq.ahmad, cstanton, Wow, this seems to be a good opportunity to get our hands on the BB AI and the MayerMakes' Add-On Cape.

Also, this terminator skull by mayermakes is really a motivation to think out of the box for any possibilities to use a given module.

For learning and understanding the capabilities of BB AI, I would like to transfer the project that I have done on Rpi 3 on to the BB AI not only to resolve the bottleneck issues imposed by the Rpi, but also to add more scalability to my project.

The project develops a smart device to detect a person with the help of image processing and take voice control commands from that user. This is to be deployed in industrial and emergency assist drones. The project is developed as part of my Bachelor thesis work and as part of a Hackster Competition where we won one of the grand prizes for a project that uses a part from this concept.  We have published 2 papers regarding this project at an International conference and also in a journal.

The real need to switch to BB AI is the fact that Rpi 3 that is being used now shows a lot of lag in the performance (currently running on OpenCV 3.2.0). It also restricts the scalability of the project as it limits the AI applications that can be run onboard. So I believe, after understanding the specs of BB AI, thanks to the EVE's, it will be able to perform a lot better and faster than the Rpi and also enable scalability to my application.

Regarding the voice control, though the Rpi ran on Sphinx base for voice recognition, with BB AI, I will be able to use more advanced AI algorithms (Eg Using Tensorflow, still learning this topic as part of my ML research) to run the detection algorithm with an increase in the performance, thanks to the DSP onboard for audio analysis.

Another issue I faced with Rpi was the controlling of onboard servos on the drone (for the pan-tilt mechanism of the camera). The issue can be resolved using a relevant Rpi HAT, but with the cost of adding more electronics and thereby more weight and complexity to the drone. With the BB AI, I would like to investigate the possibility of smooth and precise control of the servo motors using the PRU's.

The availability of onboard networking interface (WiFi) and onboard storage (eMMC Flash) is an added advantage to my application as the drone will be controlled over WiFi for short-range operations. The possibility of an LTE interface for long-range operation is under discussion.

Regarding my technical experience, I am a Computer Science graduate and I am passionate about Robotics. Currently learning on the theory and practices of Machine learning. In case I am selected to receive one of the kits, I will be documenting my project's transition from Rpi to BB AI and the further journey in the Beaglebone community space at Element14.

Jiss

Very cool project you have there! For sure the BB-Ai needs some things to be worked out, to make all your features purely run on its own hardware, as senn in my video. but it should be more than capable to do this.

current results with the EVE/DSP cores seem promissing, even the frist  examples were similar to the RPI3B+ performance we have seen earlier.
The PRUs proved perfect for realtime Servo/motor control as seen in projects like the pocketbeaglepowered Drone (found on beagleboard.org) the Replicape(a 3d printer board based on the BBB) and various other robotics projects.
I'm certain you can make it work on this Plattform. if you need Long range connectivity over LTE you may want to wait for the latest offerings of Arduino, they may add to their MKR range.

I hope you have applied for the BB-AI Roadtest and Vision thing contest!

mayermakes , thank you for your insight. Have you ever tried on interfacing any other capes? I came to know from the documentation that BB-AI would be compatible with the existing BBB capes.

Regarding LTE, since i want to implement this on drone, adding an extra companion computer would seem not viable as there are weight constraints. My plan was to interface a huawei LTE usb modem since their drivers are compatible with linux.

I personally have not reid other capes yet. according to the Posts regarding these, they are currently working on compability, the device tree overlays have to be rewritten to make them compatible.
I have high hopes that they will resolve compatibility issues pretty fast.

On LTE and MCU with LTE connctivity might be lighter than the usb stick but for sure the stick with existing linux drivers might be easier to use. Be aware that the Debian version that runsd on the BB-AI is not vanilla debian therefore the drivers might not be in the repos but have to be manually isntalled.

I personally have not reid other capes yet. according to the Posts regarding these, they are currently working on compability, the device tree overlays have to be rewritten to make them compatible.
I have high hopes that they will resolve compatibility issues pretty fast.

On LTE and MCU with LTE connctivity might be lighter than the usb stick but for sure the stick with existing linux drivers might be easier to use. Be aware that the Debian version that runsd on the BB-AI is not vanilla debian therefore the drivers might not be in the repos but have to be manually isntalled.