Publications |

Learning to Prune Branches in Modern Tree-Fruit Orchards

Wed, 01 Jan 2025 00:00:00 +0000

Watch the Demo

See our robotic pruning system in action:

Abstract Details

Dormant tree pruning is labor intensive but essential to maintaining modern highly-productive fruit orchards. In this work we present a closed-loop visuomotor controller for robotic pruning. The controller guides the cutter through a cluttered tree environment to reach a specified cut point and ensures the cutters are perpendicular to the branch. We train the controller using a novel orchard simulation that captures the geometric distribution of branches in a target apple orchard configuration. Unlike traditional methods requiring full 3D reconstruction, our controller uses just optical flow images from a wrist-mounted camera. We deploy our learned policy in simulation and the real-world for an example V-Trellis envy tree with zero-shot transfer, achieving a ~30% success rate - approximately half the performance of an oracle planner.

A Dataset for Semantic and Instance Segmentation of Modern Fruit Orchards

Mon, 01 Jan 2024 00:00:00 +0000

Diagram Placeholder

Please provide a diagram or sample images showing the synthetic data generation and automatic labeling using SAM2 and YOLO.

Placeholder: Synthetic Data Generation Pipeline

Integrating Stakeholder Perspectives into Robot Pruning Designs

Mon, 01 Jan 2024 00:00:00 +0000

Image Placeholder

Please provide an image or diagram related to the stakeholder perspectives in robot pruning designs.

Placeholder: Stakeholder Integration

Visuomotor Robotic Pruning in Planar Orchards Using Hybrid Reinforcement Learning

Mon, 01 Jan 2024 00:00:00 +0000

Architecture Diagram Placeholder

Please provide an architecture diagram showing the hybrid RL pipeline (PPO, Stable Baseline 3, SKRL).

Placeholder: Visuomotor Robotic Pruning Pipeline

SRTGAN: Triplet Loss based Generative Adversarial Network for Real-World Super-Resolution

Sat, 01 Jan 2022 00:00:00 +0000

Diagram Placeholder

Please provide a diagram or sample images illustrating the SRTGAN pipeline.

Placeholder: SRTGAN Architecture

DSLR : Dynamic to Static LiDAR scan Reconstruction using adversarially trained autoencoder

Tue, 01 Sep 2020 00:00:00 +0000

Diagram Placeholder

Please provide a diagram or sample images illustrating the DSLR (Dynamic to Static LiDAR scan Reconstruction) pipeline.

Placeholder: DSLR Architecture