ProGuard: A Real-Time Preemptive Artificial Intelligence-Based Anti-Pinch System for Bus Entryways on Low-Power Edge Devices Kettering University

Format:

Book

Conference/Event

Author/Creator:

Bradley, Hudson, author.

Zadeh, Mehrdad, author.

Tan, Teik-Khoon, author.

Conference Name:

WCX SAE World Congress Experience (2026-04-14 : Detroit, Michigan, United States)

Language:

English

Subjects (All):

Buses.

Performance tests.

Sensors and actuators.

Hardware.

Artificial intelligence (AI).

Research and development.

Cameras.

Frames.

Children.

Local Subjects:

Buses.

Performance tests.

Sensors and actuators.

Hardware.

Artificial intelligence (AI).

Research and development.

Cameras.

Frames.

Children.

Physical Description:

1 online resource

Place of Publication:

Warrendale, PA SAE International 2026

Summary:

This paper proposes ProGuard, a novel approach to preemptive pinch detection systems for buses. ProGuard utilizes state-of-the-art AI object detection algorithms to identify potential pinching events in bus entryways before pinching occurs. Modern conventional anti-pinch systems, such as pressure sensors or hall effect sensors, often rely on mechanical contact before triggering. While these systems are established safety mechanisms, they are reactive and therefore require some level of pinching before triggering. This reactive approach presents numerous safety concerns for passengers, especially when considering children on school buses. Existing preemptive detection methods, such as infrared or ultrasonic sensors, solve the problems presented by these reactive detection systems. However, these systems either lack the range or environmental resilience needed for reliable operation in buses. The critical nature of anti-pinch systems requires a robust and reliable solution that can adapt to various applications and environments. Our study investigates an AI-based approach that leverages the YOLOv11 nano object detection model to detect people and backpacks in real-time. We performed a comparative study on various model formats to find the best-performing format on the chosen edge compute hardware. Our experimental results revealed that when using the IMX model format on an AI-accelerated camera, ProGuard can achieve 24 frames per second and an inference time of 125ms while running on a Raspberry Pi computer. Performance tests on this model showed a mAP@0.5-0.95 of 0.522, putting ProGuard on par with baseline YOLOv11 nano performance. These results demonstrate that ProGuard offers an efficient and real-time alternative to current pinch detection approaches while operating on low-cost consumer hardware

Notes:

Vendor supplied data

Access Restriction:

Restricted for use by site license