Intelligent bugs mapping and wiping (ibmw) robotic system for precision pest management in almond orchards

Overview

The Intelligent Bugs Mapping and Wiping (iBMW) system is an autonomous robotic platform designed to monitor and manage Navel Orangeworm (NOW) populations in almond orchards. NOW is one of the most damaging pests in almond production, and current management practices rely on broad, calendar-based pesticide applications that increase costs, raise environmental concerns, and often treat areas that do not need intervention. The iBMW replaces this blanket approach with precision scouting and targeted spraying, enabling growers to apply treatment only where pest pressure is confirmed.

The system integrates pheromone traps mounted on an adjustable rod actuator, high-resolution imaging, and deep learning algorithms to recognize and count NOW moths in the field. By mapping infestations both spatially and temporally, the iBMW helps growers and crop advisors make data-driven decisions about treatment timing, location, and intensity. The result is reduced chemical input, lower operational costs, and improved sustainability for almond production.

Technical specifications

Key features:

• Adjustable pheromone trap actuator that raises traps to six or eight feet to capture moths across the full canopy profile
• High-resolution onboard cameras for image capture under field conditions
• Deep learning-based recognition and classification of Navel Orangeworm moths, trained on field-collected video and image datasets
• GPS-independent navigation using deep learning to adapt to real orchard environments rather than relying solely on geolocation
• Variable-height spraying module capable of site-specific, precision application of pest control treatments
• Emergency shut-off system to prevent injury to workers, crops, or equipment if unexpected obstacles appear
• Operating speed of 3.7 miles per hour, allowing efficient coverage of commercial orchard acreage
• Edge computing on NVIDIA Jetson TX2 for real-time inference and decision-making in the field

How it works:

The iBMW moves through the orchard while the pheromone traps attract NOW moths. High-resolution cameras capture images of the traps, and deep learning neural networks running on the onboard Jetson TX2 processor identify and count moths. The system generates a spatial and temporal map of pest pressure, identifies high-risk zones, and triggers the sprayer module to deliver targeted treatment only where needed.

Technology readiness level

The iBMW is currently in active development and field validation. The core cognition mechanism, including deep learning-based NOW recognition, has been demonstrated on the Jetson TX2 platform. The team is conducting ongoing field trials in a commercial almond orchard in Merced, California, planted with Nonpareil, Carmel, and Monterey varieties on Lovell peach rootstock. Multiple field visits have been completed to collect video and image data for training and refining the deep learning models. Current work focuses on optimizing recognition accuracy, validating the precision spraying module, and demonstrating end-to-end mapping and treatment workflows under real orchard conditions. The system is positioned for continued pilot-scale validation before broader commercial deployment.