***Project Cerberus - The Deadman's Smartwatch*** ***The Problem: As XKCD puts it, you can have millions of dollars of supercomputers at your disposal, but the most effective solution to cracking a password is a $5 wrench on the operator. If you know your passwords, you can be made to divulge them. Worse yet, odds are if you know your password they aren't horribly as secure as they could be. If you were to have a device that manages your cryptographic keys and passwords, that's better, especially if it generates it, and keeps them in a secure fashion. So, let's say you've got a device that can generate and store passwords securely. Great! So now what? It can still be taken off of you, and is likely still protected by just a password. This is where something like the deadman's smartwatch (Cerberus, after the Greek mythos of the watchdog of Hades) comes in. You've got all the features of a smartwatch, and a password fob. Add in the biometric monitoring (PPG pulse sensing, closure sense on the buckle, and EKG on the wrist strap [assuming this project gets to the point where the rather pricey option of Flex PCBs can be budgeted in]), and you have a device that is locked to your wrist, tracking your vitals. With some clever programming, it might even be possible to track trends, and determine physiological state (although this could be very hit or miss) and whether or not someone may be in distress. ***The solution: Project Cerberus. A vital tracking smart watch. Take it off? It wipes your keys. It loses your pulse? There goes the keys. A device like this could also be used to securely transport files, linked to the runner, kind of like Jonny Mnemonic. Very cyberpunk-y. Interfacing with other devices could be through the BTLE (most modern phones and computers), NFC (many phones, some USB readers), or through displaying the code on screen (QR or text) for manual entry. This will involve both the firmware for the watch, as well as an app (Android first) and perhaps integration with a computer (likely with a custom USB device and some sort of Linux userspace driver). There certainly are multiple challenges to overcome (actually creating the hardware for one, and funding for the last revision), and a great many skills I'll need to learn (I've not done electronics miniturization like this project will require, or the full stack design needed, from case and PCB working together, through to the applications required on the computers/phones to make it useful). ***The plan: Iteration 1: laid out development board, including on board debugger/programmer (using CMSIS-DAP for the nRF52), power measurement, and plenty of test points. Iteration 2: fitting it into a watch sized form factor, dropping all the debugging and extraneous test points, and usable as a watch. I'm also planning on building a number of beta units and have a few testers run them through their paces. Iteration 3: (Funding dependent). Using higher layer count PCBs (6 layer most likely required), FlexPCBs, and lots of BGA and chipscale packages (the CSP variant of the nRF52 isn't even scheduled to be available until later this summer anyways). This stage is just short of being production ready.