The main attraction this week was the first milestone for our group. It was the first time we could gather some feedback on our progress from supervisors and CHIC alumni.
On the engineering side, we wanted to have a small demo, at least some code to communicate with our sensor. This was complicated by a delay in the order process which meant our microcontroller would show up on Tuesday, with the milestone on Friday! To mitigate this, we decided to switch to another development kit for the demo. It allowed us to start earlier on the software, but it also means we will have to port our code to the real chip at some point. Speaking of this, maybe it is time to write a bit more about the technologies we will use in our product.
At the core of the system is a Nordic Semiconductor nrf52. This chip combines a microcontroller and a Bluetooth Low Energy (BLE) interface in a single part, which reduces the overall size of the circuit. As everything will have to fit in a wearable, this is a huge advantage for us. It is also compatible with the cheaper nrf51 series: we will be able to optimize our energy consumption and costs without starting from scratch.
On the software side, we will Apache Mynewt, a software package combining a real-time OS, hardware drivers and, perhaps most important for us, a BLE stack. It has already widely been deployed on IoT products (authors claim millions of devices), which is a good sign in term of stability and ability to deliver products.
All our software stack is Open Source so far, and this is not an accident. Aside from the ethical questions, open source has several advantages; First, we don’t have to pay expensive licenses to use those products. But more importantly, Open Source reduces our risks: if we find a bug, we can fix it without waiting for the vendor. In fact, our first patch to Mynewt was accepted last week.
That’s it for the engineering details, see you next week!