.svg)
GlobalSense is a San Diego-based technology startup pioneering a new approach to automotive diagnostics and event detection. Founded by co-founder and CEO Dr. Saeid Safavi in August 2021, GlobalSense is transforming how sound can be utilized as a powerful tool for extracting deep significance from generally common events. Its initial endeavor is to fill the significant gaps left by traditional automotive diagnostic tools, developing an innovative solution for fault diagnostics and collision detection using a new onboard device powered by audio-detection algorithms.
“Our focus is on sound sensing, or acoustic sensing, which we believe is a new sensory mechanism that can be introduced to many industries," says Safavi.
With engine diagnostics, technicians can perform some fault detection through built-in sensors using the ODB protocol, but oftentimes trained mechanics simply spot a fault, such as a piston misfire or engine knock, solely through the changed sound that the automobile produces. Audio detection aspect is an opportunity to unlock new levels of fault detection, and GlobalSense aims to embrace that while improving it through the precision of technology.
Similarly, built-in automotive event (accident, car break-in, etc.) detection relies heavily on accelerometer-based sensors, but commonly these systems have a too-high detection-failure rate in certain types of incidents. Safavi explains, "These devices are all based on acceleration and deceleration as the only sensory mechanism, and they miss 30% of collisions today simply because some collisions do not follow the algorithm for the threshold that they set for, like car rollovers or sideswipes."
To create a solution that could accurately and reliably detect faults, GlobalSense needed to access vast amounts of high-quality sound data, which it found in a partnership with an automotive auction inspection company and created relationships with a few auto auctions that send the startup daily audio recordings from their fleets of thousands of cars. Then they needed a platform that could use that data to train and update specific detection models and accommodate real-time data processing, enabling them to improve the accuracy of their models quickly.
To build their acoustic detection tool, GlobalSense leveraged artificial intelligence (AI),which can be trained to learn specific acoustic features of various events or component faults. Due to the mobile nature of automobiles and the rapid latency needed for their specific types of events, the company chose to focus on edge AI processing, meaning that the sound-detection models would run directly on the device itself, rather than being sent to a cloud-computing datacenter for processing before the results . Edge AI is also essential for preserving privacy by ensuring that only relevant information (such as events or faults) is detected locally on the artificial intelligence platform. This approach eliminates the need to record or transmit any raw data, which could otherwise contain sensitive information like human conversations or other background sounds.
Edge Impulse is the leading edge AI platform, offering a comprehensive workflow for creating edge AI algorithms, so after some initial attempts to build their own AI sequences, GlobalSense chose to employ Edge Impulse to accelerate their development of the necessary sound-based diagnostic models. The integration allowed GlobalSense to refine its models more quickly and accurately than their initial attempts.
"Before Edge Impulse came to play, you had to do everything in-house,” says Safavi. “The cycle from training the data and putting it in the hardware and testing it on the hardware was much longer."
GlobalSense leveraged Edge Impulse’s platform for a streamlined, iterative approach to model development and deployment. As data from auctioned vehicles arrived daily, GlobalSense uploaded it to Edge Impulse’s platform to test their models, adjusting and refining them based on real-world data. Edge Impulse’s easy-to-use interface and robust features allowed GlobalSense’s team to quickly identify and correct issues, improving the efficiency of their model training process significantly.
GlobalSense leverages a range of AWS services and resources, including EC2, RDS, S3, container services, Kubernetes, and Elastic Load Balancing. These services support diverse functions, from data collection and storage to enabling services for current customers, such as auto auctions.
Alongside Edge Impulse, GlobalSense is working with key partners in the technology and automotive sector to enhance data collection and labeling. A partnership with Syntiant, an AI-focused chip manufacturer, provided GlobalSense with specialized hardware for processing sound data, helping to improve the energy footprint of their models.
"We've been working with Syntiant for over two and a half years, integrating their chip into our customer product with a commitment to 2.5 million chips over time,” says Safavi. “It’s a mutually beneficial partnership — Syntiant gains from our customer base, and we leverage their technology and resources to enhance our offerings."
“The Edge Impulse toolchain has helped us shorten our time to market for some products by an order of magnitude.” —GlobalSense CEO Saeid Safavi
With the partnership, GlobalSense has been able to start assembling their customer portfolio. This includes automotive aftermarket device manufacturers, auto auctions, dealerships, and insurance companies. Now in mass production with an aftermarket device manufacturer, GlobalSense is positioned to expand beyond automotive into other fields where sound data can be used for predictive maintenance and anomaly detection.
“Edge Impulse has enabled us to potentially save several million dollars per year for one customer by helping us deploy the smallest possible models, reducing our overall BOM costs,” says Safavi. “Additionally, the Edge Impulse toolchain has helped us shorten our time to market for some products by an order of magnitude.”
By leveraging Edge Impulse’s powerful platform, GlobalSense has created a scalable, high-performance solution that fills a critical gap in the automotive industry and beyond.
Beyond automotive, GlobalSense has its eyes — or, ears — on multiple industries, including industrial IoT, HVAC, gas and oil, and more. Their groundbreaking work with sound-based sensors will bring an entirely new perspective to many fields, leveraging edge AI to make diagnostic processes more accurate, cost-effective, and scalable.
"Automotive is the point of entry for us, because we know the industry,” says Safavi. “But as we build perfection and experience in this specific form of sensory mechanism, we plan to extend it to industrial IoT, HVAC, fault detection, or other applications that have moving parts and make noises."
GlobalSense’s partnership with Edge Impulse has enabled them to build a pioneering sound-based diagnostic solution that delivers real value to the automotive industry. This collaboration highlights the potential of edge AI to transform complex, data-intensive fields and offers a glimpse into the future of sound-based predictive maintenance across sectors.
Contact us to discuss how Edge Impulse can help bring your solutions to market.
