Many of us are tracking our body activity, own sleeping, eating and exercise habits with devices like Fitbit, Apple Watch, Samsung Gear, etc. And we appreciate all the benefits that collected data is telling us to help improve our lives. Now this is not only for humans, the internet of pets has arrived. Our beloved four-legged family members can be monitored in regards to its location, activity and even emotional state or any possible health issues. In fact, pet and livestock wearable market in terms of medical diagnosis is expected to grow from 1% to 13% during 2015-2025. This market in general will hit $2.36 Billion mark by 2022, another report suggests. Investors are pouring million-dollar funds into R&D innovations at existing companies and also scout for outside start-ups who can quickly deploy fresh ideas into the market.
Recent CES events confirm the growing interest and potential in pet wearables, however, we didn’t notice any ground breaking tech introduced or blended into the presented products, most of them work pretty much like the human fitness trackers, with some extra bells and whistles here and there. However, newly introduced devices tend to be more personalized, and they come with beautiful industrial design in most cases.
Technology segmentation in pet wearables (according to Grand View Research):
- GPS (real-time positioning).
- Sensors (accelerometer/gyroscope to detect movement, temperature-based activity sensors for monitoring health-related biometric parameters, etc.).
- RFID (stores data using electromagnetic forces for power, and communicates that data to a device that interprets it).
Pet wearable application outlook:
- Identification & tracking
- Behavior monitoring & control
- Facilitation, safety & security
- Medical diagnosis & treatment.
Where is the limit for all of this? Ian Pearson, a futurologist, says AI program such as Amazon’s Alexa will emerge soon and it will help to get to know your pet’s sounds and “translate” them into basic statements, such as telling you they’re hungry or in pain – though he points out many owners can already read these noises. He pictures devices that will remotely take your dog for a walk for you, using commands.
Pet wearables industry boom has also touched us here at Lemberg Solutions. The start-up that we are going to talk about - Horse Analytics - is not related to dogs or cats, which are the most popular pets out there. We will talk about the horses this time. These noble animals are pretty expensive and not everyone can afford one, let’s face it. On top of that you have to have a whole team around it: horse coach, horse rider, stables keeper, horse doctor, etc. Now you realize that this is not so simple anymore. Of course animals can’t talk but you and your team can understand them a lot better when you monitor your pet 24/7 and collect data about its activity, exercise, sleep and behavior. So how do you make it happen? You can imagine how exciting it is to work on a challenge like this. Keep in mind, our project is not just a mere tracking device with a bunch of sensors built it, the software that we are building comes with a self-made Machine Learning (ML) algorithm, but we will come back to this a little later.
Back to the hardware, Lemberg Solutions team had to work on a device that is supposed to include multiple sensors and components: accelerometer, gyroscope, BLE module, data storage, GSM module, heart rate sensor. Because of several hardware limitations we decided together with the product owner to start building an Android-based hardware emulator and use basically a regular phone on the horse that would work and collect data in real time like the actual device would. Of course we had to be mindful and design the right architecture which also included the cloud with ML algorithm and beta-tester mobile app as well as user mobile apps.
The most intriguing part was related to the ML algorithm. The goal is to detect precisely if the horse was walking, or trotting, or galloping, etc., and if it was on the left or right side. We understood that this is a key challenge of this project, and that having this algorithm in place would ultimately make it a much better product with greater value. Our Data Scientist proposed a development approach which was the optimal option to go with and we opened an account in MS Azure IoT suite to kick off the process.
As it typically happens in agile development, along the way we also came up with the need to add new interesting features, like the Dressage navigation. For those of you who are a bit unfamiliar with equestrian sports, Dressage is the highest expression of horses training where the horse and rider are expected to perform from memory a series of predefined movements around the arena. So what’s so special about it? Well, for one thing we had to include text-to-speech engine that would allow for the horse rider app to read out all the exercise commands in the right sequence and speed. And the rider can hear these commands in his ear piece. The benefit is that the coach doesn’t have to shout out across the arena to the rider any more. Secondly, we had to brainstorm and come up with the best solution for Dressage navigation, and it was the beacons. As we know Estimote beacons give 1 meter accuracy, so it was a pretty obvious choice. Check out how exactly we have implemented this feature in the Dressage feature case study.
This project is in the phase of ongoing development, leveraging our IoT and embedded development services. We look forward to delivering this product to the customer as soon as possible. And who knows what other new challenging features would come into sight down the road. Anyway, If you have a project idea in mind, but don't know where to start, we're always here to help you to provide high-quality IoT development services.
