Industrial IoT

The Internet of Things (IoT) is rapidly connecting and bringing the physical world online. This paradigm shift is essential to businesses, transforming everyday objects and critical assets into smart, interconnected systems by leveraging advances in sensors, analytics, models, and AI.

PARC is at the forefront of this rapidly changing technology landscape. With IoT as a key focus area and innovation pillar, we are working with clients and partners to develop the technologies that can enable this mass transition and help organizations gain strategic, competitive advantage.

 

The Internet of Things is essential to organizations across numerous industries, providing real-time awareness into how their systems really work.

With analytics front of mind, our labs are creating smart sensors and systems that enable rapid data collection at scale, which can then be used to develop actionable insights. We enable end-to-end solutions that address critical challenges such as system reliability, anomaly detection, health management, environmental monitoring, streamlined operations, and predictive maintenance.

Our comprehensive IoT technology portfolio includes:

• Field-deployable sensors and devices to monitor/manage systems
• Advanced analytics to reliably make sense of IoT data in real time
• Hybrid physics/AI models that enable a deep understanding of systems
• Adaptive optimization to aid proactive scheduling/planning
• Security/privacy to ensure system confidentiality/integrity

 

Linking the physical and digital worlds

 

 

Asset performance management & predictive maintenance

In transportation systems, critical infrastructure environments, and factories, unexpected machine failure or degradation can disrupt operations and lead to significant unplanned downtime, costs, business losses, and even injuries. Therefore, it’s important to detect potential issues in the early stages, before failure occurs, to allow proactive maintenance to be performed during off-peak production hours.

IoT-enabled predictive maintenance can be a key first step in the broader quest to enable self-aware, self-adaptive systems. Read about PARC’s engagements with East Japan Railway (JR-E), VicTrack, and other case studies to learn how we are pushing the frontiers of science and technology to provide advanced commercial IoT solutions.

 

Ocean drifters gather maritime environmental information

As another example, PARC’s Ocean of Things project is deploying small, low-cost ocean drifters to collect environmental and human impact data including sea surface temperature and activities, and even information about marine life moving through the area. Each solar-powered drifter is made of environmentally safe materials and has a suite of onboard sensors including a camera, GPS, microphone, hydrophone, and accelerometer. Advanced analytics process and share the data gathered, enabling persistent maritime situational awareness. Data about changes in the ocean environment can be used in a broad array of areas including ocean pollution, aquafarming, and transportation routes. We can now can gather data that we could never track before.

We make sense of the data

IoT solutions typically must address three big issues: 1) deploying robust, low-cost sensors for quality data; 2) the 3Vs of data (volume, velocity, and variety); and 3) translating data into effective decisions towards making a business case for adoption.

For traditional AI algorithms and deep learning, huge amounts of labeled training data is necessary. The cost and expertise needed to label data at high volumes can be overwhelming. Accuracy rates, data reliability, and computational resources can impact adoption. In addition, disparate data sources create a challenge that needs to be considered by edge/cloud algorithms.

Over the past 50 years, our expert team has pioneered ground-breaking research in different areas that are now coming together to allow IoT-enabled digital disruption. Our approach combines advanced sensors with secure edge/cloud architectures to reliably monitor the phenomena of interest, and hybrid physics/AI models that represent those phenomena without needing extensive labeled data sets. Anomaly detection algorithms detect subtle changes from nominal behaviors in complex, multi-modal data. Model-based diagnostics and prognostics analyze and predict issues, and actionable recommendations mitigate and manage the IoT systems. Together, these capabilities result in highly effective end-to-end IoT solutions that maximize system performance and create compelling new digital business opportunities.