"The original paradigm of simply moving all of the data generated on the edge to the cloud for analysis and machine learning has run into three fundamental barriers, namely privacy, reliability and latency. Running these algorithms efficiently on edge eliminates these barriers and allows AIoT products to scale much more rapidly", says Sree Harsha Angara, guest of our new episode.

In this new episode we discuss,

• the development of AIoT and Machine Learning,
• fundamental barriers and challenges and
• how the new ModusToolbox^TM ML supports here.

More information about ModusToolbox^TM ML. Register now for the webinar "Unlocking AI for IoT Devices with ModusToolbox™ Machine Learning" on Thursday 22 July, 9:00 AM PDT / 12:00 PM EDT.

Guest: Sree Harsha Angara
Date of publication: 16 June 2021

Moderator:

The potential of the Internet of Things is well known. But how do we actually implement it? How can people and companies benefit from it? In this podcast, I meet with experts from infineon, partners and customers who tell me how it can work and what it takes to Make IoT work.
My name is Thomas Reinhardt, I am your host, and I am excited to bring this podcast to all of you.

The number of connected devices keeps going up as consumers’ intent to increase their level of comfort and convenience, while enterprises intend to create new value and revenue streams through new service offerings.

As we look ahead, more capable smart devices are going to be available, shaped by different key trends. One of these key trends is the combination of AI and IoT at the edge, known as the Artificial Intelligence of Things (AIoT), which provides machine learning capabilities in connected devices. Today, I talk with Sree Harsha Angara about AIoT, machine learning and Infineon’s new ModusToolbox™ Machine Learning.
Welcome Sree Harsha, it’s a pleasure having you as guest today.

Guest:

The pleasure is mine. Thanks for inviting me.

Moderator:  

In the last episode, I talked with Vikram Gupta about how AIoT is trending and having a huge impact on IoT. But what does AIoT mean and what are the challenges of developing innovative, smart, connected and secured IoT devices of the future - especially when it comes to AIoT?

Guest:

Machine Learning is progressing at a rapid rate with new Deep Learning algorithms able to solve historically difficult problems using data-driven design principles. This is especially exciting in IoT where the rapid increase in connected devices has led to an explosion in the amount of data being generated at the edge.

These new algorithms are playing a critical role in advancing the next phase of the IoT revolution. This combination of AI and IoT, known as the Artificial Intelligence of Things (AIoT), provides machine learning capabilities in connected devices, enabling them to perform intelligent tasks. According to Markets and Markets, the AIoT market is expected to increase from US$5.1 billion in 2019 to US$16.2 billion by 2024, growing at a CAGR of 26 percent.

The original paradigm of simply moving all of the data generated on the edge to the cloud for analysis and machine learning has run into three fundamental barriers, namely privacy, reliability and latency . The natural answer people have arrived on is to shift ML algorithms which typically run on the cloud at the edge.

Moderator:

You talked about three barriers: privacy, reliability and latency. I mean, latency seems to be clear to me. However, could you please elaborate a bit on these barriers?

Guest:

Certainly, a great example to think about are voice-based smart assistants. Firstly, when you interact with an assistant, the time it takes to make a round-trip to get answer is generally a poor user experience as that’s not the natural way to interact for humans. Secondly, reliability and bandwidth of your internet connection is also critical when you think about these assistants run on wearable devices such as smart watches: You don’t always have a perfect, reliable connection to the cloud. Thirdly, with the proliferation of these assistants everywhere privacy is always top-of-mind and trusting service providers with sensitive voice data is always a challenge.

Running these algorithms efficiently on edge eliminates these barriers and allows AIoT products to scale much more rapidly.
Now, one of the big challenges in developing ML algorithms for AIoT products is that they traverse various technical teams, ranging from data scientists to embedded firmware developers. The typical development tools and workflows used by data scientists for developing these algorithms are not geared towards IoT and the task of integrating these workflows efficiently into the larger embedded software stack requires significant effort from firmware developers.

Moderator:

That sounds really challenging. But how we as Infineon help to reduce the time to market and the expense required when launching high-value, high-quality devices of the future onto the market?

Guest:

For many years we have been offering our customers ModusToolbox™, an intuitive and easy-to-use collection of libraries, software and tools, compatible across environments (Linux, macOS, Windows), that provides software developers with an immersive, secure, and comprehensive experience to build embedded and IoT products. It offers a modern software development approach based on an open source system with pre-built tools and seamless integration into third party applications, allowing developers to use the tools that they want and build like the best. It also leverages popular IoT focused development frameworks such as Mbed OS, Amazon FreeRTOS, AliOS Things and Zephyr. Together, they enable a comprehensive development experience for customers creating converged MCU and Wireless systems.

Moderator:

And what is the actual benefit of that?

Guest:

ModusToolbox™ considerably simplifies the development of IoT products which use Wi-Fi and Bluetooth/Bluetooth Low Energy IoT products in combination with RTOS system microcontrollers, such as those from the PSoC™ family. Developers can use the integrated middleware and code examples to easily connect their IoT products with leading cloud software platforms or proprietary cloud services.

ModusToolbox™ also contains solutions to support popular ecosystems and cloud management tools like Pelion Cloud Management and MbedTM OS, Amazon Web Services (AWS) IoT and FreeRTOSTM SDK as well as Infineon AnyCloud IoT.
In addition, it provides specific tools like the low-power assistant, multi-radio smart coexistence, secure authentication, and over-the-air updates reduce the time and expense required when launching high-value, high-quality products onto the market.

So – you can see that Infineon is already providing flexible, easy-to-use tools and solutions for creating smart devices of the future

Moderator:

And within this ModusToolbox™ we offer now a new feature for AIoT?

Guest:

Right. ModusToolbox™ Machine Learning is a new feature in ModusToolbox Software and Tools. ModusToolbox™ Machine Learning (ML) enables deep learning-based workloads on Infineon’s PSoC™ microcontrollers (MCUs). It provides middleware, software libraries and special tools for designers to evaluate and deploy deep learning-based ML models. This feature allows seamless integration with existing frameworks available in ModusToolbox so that ML workloads can be integrated into secured AIoT systems.

Moderator:

So with ModusToolbox™ ML we support to make now the AIoT work?

Guest:

Yes, and this brings a lot of benefits to the customers, primarily focusing on bridging the gap between data scientists and embedded developers by providing a unified software flow.

ModusToolbox ML allows developers to use their preferred deep learning framework, such as TensorFlow, and deploy them directly to PSoC MCUs . In addition, the tooling helps designers by optimizing the model for embedded platforms by reducing the size and complexity using a variety of techniques such as quantization.

One of the other key features this toolset brings is helping visualize how these optimization techniques impact model performance so you can make the right trade-off’s between performance vs the size/complexity of running the model efficiently on a PSoC MCU.
To help getting started quickly, we also provide code examples and IoT-focused development kits to have a smooth developer experience that reduces the complexities system developers face when developing AIoT applications. These typically require a seamless Machine Learning workload integration, along with compute, connectivity and cloud domains. ModusToolbox ML can address this by giving developers the ability to simply able drop this functionality into any existing cloud or connectivity examples today.

So with our ModusToolbox™ ML our customers can expect great features that will help them to make AIoT work. And a positive user experience is what fuels us to continuously enhance our products and solutions with every version and feature that we release.

Moderator:

Great! Seems as if ModusToolbox was built to make our customers’ life easier and more efficient by removing development barriers and allowing to deliver quality products to market faster. So if our listeners now want to see it for themselves: What do they have to do if they want to use ModusToolbox?

Guest:

You can find the links to download ModusToolbox or to learn more about the ML solution in the description of the podcast or its transcript.
•    View code examples in the ModusToolbox GitHub Repository
•    Join the Cypress Developer Community to get access to online documentation, online videos, regular live developer trainings

Moderator:

Thank you very much. This brings us to the end of this episode. Thank you very much, Sree Harsha, for your exciting insights. I'm already looking forward to seeing more self-learning IoT devices in the markets to make the IoT work.

Dear listeners, for more information, please visit infineon.io. We will publish the next episode soon. Take care. And until the next time. Take care.