Asthma is a major health concern that worsens with each passing year. One in 12 people suffers from it — up from about one in 14 nearly 20 years ago — and half that people who have asthma suffer an attack in a given year. Asthma is particularly hard on children: it’s the top reason for missed school days and the leading chronic disease among children.
According to the Centers for Disease Control, more than 3,600 people died in the United States in 2015 from asthma; the World Health Organization puts the global number at approximately 250,000 deaths per year. Many of these deaths are avoidable.
These kinds of statistics motivated the founders of Amiko to attack the problem with digital technology. Amiko’s Respiro smart inhaler sensor is bundled with an app that the patient installs on his or her smartphone. The AI-enabled sensor collects inhaler use data without disrupting the medication delivery pathway, and sends data and private user feedback to the app via Bluetooth Low Energy.
Where required, a professional dashboard offers clinicians a suite of applications, remote monitoring tools, artificial intelligence-enabled therapy suggestions, and data analytics. This information helps them coach patients and offer support, while direct care givers can see the actual dose and its efficacy.
The Respiro uses machine learning (ML) to interpret vibration data from the inhaler. The sensor is trained to recognize the patient’s breathing pattern and inhalation time, and can calculate important parameters such as lung capacity and inhalation technique. The modules, powered by Arm Cortex-M microprocessors, use ML to track and report on inhalation technique, flow rate, volume and other parameters.
The Respiro’s designer chose a device like the Cortex-M to optimize for lower-power processing and footprint, and the overall design was so innovative it won first prize at the IBM Watson AI XPRIZE’s annual Milestone Awards in December 2017. Respiro sensors are compatible with market-leading inhalers, such as NEXThaler® (Chiesi Farmaceutici), Ellipta® (GlaxoSmithKline) and Spiromax® (Teva Pharmaceutical Industries).
This is just one example of not only advancing applications through machine learning but leveraging and better compute power within edge devices to improve responsiveness, security and the user experience.
It’s also an example of the kind of innovation in AI/machine learning and edge computing that can be found in this year’s Arm TechCon, which is being held Oct. 16–18, 2018, in the San Jose Convention Center.
There’s an entire track on edge computing and machine learning waiting for you. Consider a just a few of the presentations on tap.
- Unique, hands-on workshops focusing on AI and ML: Machine learning on Arm Cortex-M microcontrollers and accelerating and optimizing machine learning on Arm Cortex-A.
- Tirias Research’s Jim McGregor will weigh on the options available to developers, because there is no one-size-fits-all solution in the space.
- Shawn Prestridge from IAR systems will discuss how to optimize IoT nodes for machine learning.
- One you capture data at the edge how do properly analyze it to help transform your business? Arm’s Stephen Barton discusses how to analyze machine learning inference with Arm tools.
These are just four of 18 Arm TechCon sessions devoted to edge computing and machine learning –the most-packed track of any of the eight Arm is highlighting at the event. Other topics include automotive, system design methodology, embedded software development and trust and security.
As the pace of design quickens in the IoT era, it’s more important than ever to stay current, so come join us this fall. If you need more reasons, check out our justification toolkit. There are resources including sample letters to your manager to help you come join us in San Jose.
Register now and get discounted rates. See you in October!