Most personal safety technology is reactive. Cameras record incidents. Phones let you call for help. Trackers show where you were. But none of them necessarily give you time to react before something happens.
SIMS is taking a different approach. Using on-device AI vision, the wearable alerts users when a person or object is approaching from behind, turning situational awareness into a proactive capability rather than a forensic one. The device was showcased at CES as a partner demonstration running on Alif Semiconductor® silicon, providing a practical example of wearable edge AI applied to a real-world safety challenge.
SIMS founder Sean Siembab has described the idea as stemming from a simple question after being struck from behind by a cyclist while wearing headphones: why can cameras record incidents, but not warn people before they happen? That question ultimately led to a product designed for runners, walkers, cyclists, security personnel, and others who may benefit from greater awareness of their surroundings without constantly looking over their shoulders. In our CES demonstration, the same broad applicability was highlighted, including use cases ranging from nurses to schoolchildren.
What makes the design especially compelling is the way it has been built around everyday use. SIMS says the device is designed to be worn on the back of clothing or attached to a backpack. It can alert the user through a smartphone or smartwatch, while the broader ecosystem includes a mobile application, smartwatch interface, and SOS capabilities that can notify emergency contacts and share GPS location information when needed.
Delivering that experience is fundamentally an edge AI problem. A wearable designed to warn users about something approaching from behind cannot rely on a delayed decision path. It must capture the scene, interpret what it sees, and generate an alert quickly enough for the wearer to respond. The SIMS device performs this detection locally using an Ensemble® E7 MCU with Arm Ethos-U55® NPUs to accelerate AI workloads. By moving intelligence directly onto the device, the system can respond immediately without depending on a cloud-based analytics loop.
The processor architecture plays an important role in enabling that capability. Ensemble E7 combines application processing, AI acceleration, and real-time processing resources in a single device, allowing products like SIMS to handle image capture, inference, alert generation, power management, and companion-device connectivity simultaneously. A safety wearable may be physically compact, but the workload remains demanding. Reliable real-time decision making must coexist with the power, thermal, and size constraints that users expect from wearable devices.
That balance is what makes the SIMS design noteworthy. Much of today’s wearable technology falls into one of two categories: devices that passively record information, or devices that offer intelligent features but remain heavily dependent on connectivity, smartphones, or frequent charging. SIMS is targeting a more practical middle ground. The company describes a lightweight camera-based device capable of capturing images at a high frame rate, triggering external alert sirens in SOS mode, and sending notifications and location information through its companion system when necessary. The underlying philosophy is straightforward: perform the critical interpretation locally, then use connected systems to support the user rather than replace the core decision-making process.
There is also a broader lesson here about privacy and product design. As AI becomes increasingly common in consumer electronics, users are paying closer attention to how and where their data is processed. Applications that can make decisions locally often benefit from lower latency, reduced bandwidth requirements, and greater user control over sensitive information.
Wearable cameras can quickly become problematic if they feel invasive, fragile, or overly dependent on continuous streaming. SIMS incorporates optional connected and cloud-linked features for alerts and evidence handling, while the core user experience remains centered on on-device AI detection. That distinction matters. The most effective wearable intelligence is often not the intelligence that generates the most data. It is intelligence that can make the right decision locally, at the right moment, without requiring a more complex, power-hungry, or intrusive system architecture.
SIMS also illustrates a broader shift taking place across edge AI. The most compelling products are often not those that showcase the most advanced technology, but those that apply that technology to solve a clear, immediate problem. In this case, the challenge is simple: give people greater awareness of what is happening around them. Delivering that capability reliably, in real time, and within the constraints of a wearable device is where edge AI becomes genuinely useful. And it is exactly the kind of application that demonstrates how intelligent systems can enhance everyday experiences in ways that feel natural, immediate, and practical.