top of page
Writer's pictureCarlos Gustavo Merolla

Building intelligent and reactive edge devices using the SONY Spresense

Updated: Apr 10, 2020




Introduction

This series of posts are dedicated to the Sony Spresense board (and all the related topics) and the context in which I’m using it. This is not a copy and paste tutorial, I will try to explain the concepts and details that will allow you to create amazing things in the future (that’s my hope). The entire work will be based on a sample device, an Intelligent Connected device that will be able to handle sound, pictures and other sensors. You will find some easy to understand schematics, code, etc.


The aim is to explain and put into practice concepts about multi-core programming, power management functions and sensing capabilities available on the Sony Spresense, its operating system (NuttX) and the native SDK and Arduino IDE to create reactive devices, which means, devices that are able to react accordingly to the events occurring in its surroundings trying to use the appropriated performance and power consumption levels. In addition to this “structural” type of response, these words will also take into account the content of the reaction. Under this sections, the Artificial Intelligence capabilities on the Spresense and its related tools (The Sony Neural Network Console and nnabla) will be also introduced to perform inference on images, sounds, etc.


A few words before we get started.

This work is the result of my research on equipment used for Edge applications in the context of my project, BCB, to protect Wild Life, as you could have guessed by the Logo on the upper right corner, the main goal is to protect Big Cats (Yaguaretès, a local kind of Jaguar) from vanishing from our world. While this work is focused on the tech side of it, allow me to throw a couple of lines to you to call your attention about the subject. In general terms, biodiversity is quite important for our world, we live in an ecosystem where each part plays a role, from trees to cats. Even if we don’t think about them, wild life helps to maintain the equilibrium on this planet, they are, in a way, silent house keepers of our big home which is earth.


Our unplanned progress, irresponsible use of resource is dragging wild life to disaster, they are losing their habitat, they are being killed in roads and they are even hunted for revenge or sport, they are at risk of extinction, here, in Argentina, we have just 250 Yaguaretès alive, the same is true in almost any corner of this world, we cannot watch them vanishing without a good fight, so, please, get involved, support your favorite NGO, support my project, start your own, spread the world, but, please, do something about it. Get involved.


The Context

Getting into the subject, Big Cat Brother (BCB) is composed by several devices that have to be deployed in real tough environments, there are many challenges such as the lack of electric power, standard connectivity, among other physical ones. Some devices that compose BCB are Smart cameras that can detect the presence of some subjects of interests (Humans, Cats, etc) and notify (and get notified) about those events using LoRa messages.


Other challenges are derived from the animal habits, a male Yaguareté needs about 30km 2 as his hunting territory (I say “his” because they are equals to me), which means the territory to be covered in quite big. That imposes constraints that goes from cost to easy to maintain devices. The combination of reasonable computing power to run AI algorithms (for both sound and video), low energy consumption and a reasonable price is a hard to find combination. So, let’s dig into the tech details.


 

In the following posts, we will start to discover some features about the Spresense and as we go we will find the benefits on this board and other related technologies in the context of deploying solutions quite far from the city abundant resources. Stay tuned.

 

75 views0 comments

Recent Posts

See All

Comments


  • Twitter Social Icon
  • Mail Icon
bottom of page