10:45 am

Building an Edge Device with Open Source Hardware: From Data Collection to Deployment

Sai Yamanoor - IoT Applications Engineer • Praxair

With the arrival of powerful microcontrollers and software toolsets, it is possible to run a neural network on a microcontroller. This provides a lot of opportunities in applications like image processing, keyword detection etc.

Edge Computing enables recurring costs like cellular data charges, provide alternatives when there is poor connectivity or react immediately to a critical sensor input and avoid loss of productivity, identify anomalies (e.g.: leak detection) in data streams etc. In this talk, we will demonstrate the different hardware and software toolsets available to build an edge device and walk the audience through the different steps involved in building an edge device. We will also talk about the challenges involved in building and testing an edge device

Day Two

5/14

7:30 am

Registration & Breakfast

8:00 am

Scott Austin, ESVP and Head of Country Americas • Everledger

Dave Hunt, Sales & Business Development Director • ACSIS, Inc.

Travis Wayne, Product Manager • Teklynx International

Blockchain enables companies to record every event or transaction within a supply chain. This information then can be shared in a secure, transparent and efficient manner. So what is implementing a blockchain system like? AIM has put together an expert panel that can answer that by sharing their experiences and challenges in utilizing blockchain applications with AIDC technologies, such as IoT and RFID;  and how it improved the traceability and reliability of information.

8:45 am

Artificial Intelligence: From a Buzzword to an Invaluable Manufacturing Component

Stefan Loelkes, Ph.D. - Chief Sales Officer (CSO) • MPDV USA, Inc.

Artificial intelligence (AI) has matured from a buzzword to tangible applications. It has arrived with full force in the world of manufacturing execution. In our talk, we will clarify current terminology in this field, including concrete examples where AI can be used in a valuable way to increase production efficiency. A prominent example is predictive apps that determine the quality of a workpiece or batch during production based on data from multiple smart iIoT sensors. All of that also requires proper modeling of the AI kernel; otherwise, it will not be possible to correlate historical data with current input to forecast results.

 

We will end our talk with an overview of how complete feedback loops become established in state-of-the-art production. To do this efficiently, existing IT/OT-systems are enhanced by new technology like AI. This addresses all parts of the feedback loop, beginning with the planning and scheduling of the operations. Reinforced learning algorithms can be drastically more efficient than classical heuristic planning in certain use cases. Subsequent execution and monitoring can be improved by predictive apps like predictive quality. During and after execution, advanced analytics allow the production run to be analyzed and compared to historical data. Finally, predictive algorithms help to draw conclusions from the analytics and influence the next planning cycles.

9:30 am

The IIoT challenge: Building a System Today for 2040 and Beyond.

Jim Kokal - Presodent & CEO • Wavetrix

Industrial Internet of Things (IIoT) offers different challenges than consumer IoT applications. Beyond frequently operating in harsher environmental conditions, the lifetimes of industrial products are much longer than typical consumer products.

 

Creating a product with a long life starts with examining where it is going to be used, who/what is going to interact with it, the maintenance cycle, and the expectations of lifetime. As an example, consider configuration of the device. One of the most popular options today is using an app on a smartphone to configure the product. Smart phones available in ten or twenty years will only be able to run the configuration app if it has been maintained for the entire time. That implies a hefty app maintenance cost. There may be other options that have a lower life-cycle cost.

 

Configuration is only one example, part obsolescence and connectivity issues with cellular, Wi-Fi, USB, etc. also need to be considered. Addressing these issues will affect the development, device, connectivity and support costs in both getting the service to market and maintaining it through its operational life. Creating a product to last an industrial lifetime requires forethought, ignoring these issues can radically change the business model making the venture a money-loser rather than a money-winner.

10:00 am

Networking Break & Dedicated Expo Time

10:30 am

Overview of the Hardware Development Process for IoT Devices

Saj Patel - CEO and Co-Founder • Optimal Design

Hardware design differs greatly from software development. At a fundamental level, the design and engineering process has fixed timelines for activities such as industrial design, packaging engineering, RF design, testing/validation, tooling and production ramp. These activities have to generally be completed in a linear “waterfall” fashion, which is quite different than the agile method for software. The purpose of the presentation is to provide those in the IoT industry an overview of the process, challenges, and expected milestone outcomes during hardware development. A key focus will be on presenting a process that minimized costs and maximizes outcomes. In order to design great hardware, we will highlight specific guidelines that need to be adhered to throughout the process.

This presentation will review the eight phases of hardware development in detail. In addition to the development process, data security, weatherproof/rugged design and long term reliability are key items need to be kept in mind and will be covered as a part of this talk.

11:00 am

Gaining a Competitive Edge in IoT with Machine Learning

Murali Kashaboina - Founder and Chief Data Scientist • Entrigna, Inc.

Companies are realizing that in order to maintain a competitive edge they need to derive insights from sensors and smart device data. However, many IoT solutions are doing simple tasks like predicting if a part will break. This is a great start but real-time solutions that incorporate machine learning are what will set a solution apart and lead to a competitive edge. In this presentation, you will learn about the different machine learning components and examples of how businesses could use each component to make a custom IoT solution that will lead to more accurate customer insights and increases revenues.

11:30 am

Building an Edge Device with Open Source Hardware: From Data Collection to Deployment

Sai Yamanoor - IoT Applications Engineer • Praxair

Srihari Yamanoor - President • DesignAbly

With the arrival of powerful microcontrollers and software toolsets, it is possible to run a neural network on a microcontroller. This provides a lot of opportunities in applications like image processing, keyword detection etc.

Edge Computing enables recurring costs like cellular data charges, provide alternatives when there is poor connectivity or react immediately to a critical sensor input and avoid loss of productivity, identify anomalies (e.g.: leak detection) in data streams etc. In this talk, we will demonstrate the different hardware and software toolsets available to build an edge device and walk the audience through the different steps involved in building an edge device. We will also talk about the challenges involved in building and testing an edge device.

12:00 pm

Conference Conclusion

2020 IoT Innovation North America Conference Agenda

Sign up for updates

© 2019 TWST Events

Social Media

  • Black Facebook Icon
  • Black Twitter Icon