Fast Object Detection with a Machine Learning Edge Device

TL;DR

The paper investigates real-time object detection on a low-cost edge device integrated with an embedded system for autonomous humanoid robots, comparing CPU, GPU, and Google Coral Edge TPU. It deploys a CNN-based detector with transfer learning, using TensorFlow/TensorFlow Lite and a RoboFlow dataset of 4,000 images to assess inference time, precision, and recall. The results show that the Edge TPU achieves the fastest inference (about 30 ms) compared with GPU (≈40 ms) and CPU (≈240 ms), with precision around 80% and recall below 80%, and no significant advantage from monocular versus stereo vision in this setup. The work concludes that Edge TPU–based edge devices are viable for RoboCup-style scenarios and outlines future integration with the robot’s onboard processor and additional experiments.

Abstract

This machine learning study investigates a lowcost edge device integrated with an embedded system having computer vision and resulting in an improved performance in inferencing time and precision of object detection and classification. A primary aim of this study focused on reducing inferencing time and low-power consumption and to enable an embedded device of a competition-ready autonomous humanoid robot and to support real-time object recognition, scene understanding, visual navigation, motion planning, and autonomous navigation of the robot. This study compares processors for inferencing time performance between a central processing unit (CPU), a graphical processing unit (GPU), and a tensor processing unit (TPU). CPUs, GPUs, and TPUs are all processors that can be used for machine learning tasks. Related to the aim of supporting an autonomous humanoid robot, there was an additional effort to observe whether or not there was a significant difference in using a camera having monocular vision versus stereo vision capability. TPU inference time results for this study reflect a 25% reduction in time over the GPU, and a whopping 87.5% reduction in inference time compared to the CPU. Much information in this paper is contributed to the final selection of Google's Coral brand, Edge TPU device. The Arduino Nano 33 BLE Sense Tiny ML Kit was also considered for comparison but due to initial incompatibilities and in the interest of time to complete this study, a decision was made to review the kit in a future experiment.

Figures (9)

• Figure 1: Humanoid Robots at a RoboCup Competition.
• Figure 2: A Convolutional Neural Network Components.
• Figure 3: Google Coral TPU.
• Figure 4: Precision Formula.
• Figure 5: Recall Formula.