A robotic dog that autonomously waters plants

The B-AROL-O Team are a small group of friends who work or have previously worked for an Italian company in the packaging industry. Barolo is also a wine region. The dog is called FREISA, an acronym for Four-legged Robot Ensuring Intelligent Sprinkler Automation. With this new integration of robotics, AI, and agriculture, the team says it aims to introduce a new era of smart gardening solutions.

The robotic dog works by locomoting using its four limbs. A camera module is used to observe the surrounding environment. Plants are evaluated using AI and watered by FREISA if it looks too dry. “Our robot demonstrates good abilities in identifying a target plant, navigating towards it, and examining its leaves before precisely adjusting its position to efficiently water the plant”, Gianluca Teti, member of the group, says.

Vineyard challenges

Initially, the team contemplated using vineyards as a setting for the dog. However, this choice presented certain challenges. Primarily, the vine leaves grow at a height of at least 40 cm from the ground, significantly higher than the intended height for the dog robot it planned to employ.

Also, the terrain where vines are typically cultivated, notably in Italy, tends to be uneven, posing navigation difficulties for a robot. Therefore, the team chose a more controlled environment – a small garden with tomato plants. But, the team says, this decision doesn’t rule out the possibility of continuing its project and potentially introducing the robotic dog to vineyards in the future.

The team’s objective was to assess a plant’s condition by analysing the state of its leaves and stems. The team had to collect a vast array of images and then train various networks and models.

Quadruped legged platform

The team believes that Quadruped Legged Platforms are ideally suited for a range of unstructured outdoor applications due to the ability to navigate complex terrain. They ruled out creating a robot running on wheels and tracks, noting that it would be poor at clearing obstacles over rough ground.

The group also considered a quadcopter, but felt its battery and payload would be poor. The team built the robot for less than US $500, using open source software. FREISA won the Grand Prize in the 2023 OpenCV AI Competition. The team has already identified potential improvements such as using the OAK-D Lite’s stereo camera to estimate a target’s distance.