The FEROX project aims to utilize the advancements in AI, data and robotics to improve working conditions of wild berry pickers and optimize picking operations. This blog dives into our most recent tests held in September 2024 in Rovaniemi related to the improving working conditions of berry pickers.
Berry picking is the process of harvesting berries from forests, a task that requires pickers to navigate natural terrains often lacking formal pathways. Traditionally, pickers rely on personal knowledge, experience, and simple techniques like marking trees to track their routes. This method, while straightforward, demands significant attention and mobility, which can become problematic if a picker is injured or disoriented. Pickers commonly use smartphones for navigation; however, these devices often face challenges such as poor network coverage or running out of battery, especially in remote areas with swamps or dense vegetation.
The FEROX project enhances berry picking with advanced maps, autonomous drones, and the WatchDog feature, which periodically monitors workers' safety, assists with navigation, and ensures timely support in emergencies, significantly improving safety and trust.
The WatchDog feature is a comprehensive safety system aimed at detecting and responding to emergencies in real-time. It functions by creating a mechanism for triggering emergency response through the use of GPS tracking device (i.e., GPS MiniFinder Pico) and drone technologies. The GPS MiniFinder Pico 4G allows constant sampling of the location of the picker as well as creating virtual geofences and creating an SOS alarm via a dedicated onboard panic button. The system ensures swift and effective responses for the berry pickers such as navigating themselves after being lost in dense forest, sending drones to visually inspect pickers during medical emergencies, as well as monitor work in restricted areas (e.g., breaching a hazardous area).
The GPS MiniFinder Pico 4G is used as the main trigger to initiate the emergency response.
The core features of the WatchDog are:
SOS Button: Pickers can alert a designated contact via mobile network communication in emergency situations, such as getting lost or injured.
Virtual Geofencing: Setting up defined safe zones or defined hazardous zones to help pickers stay safe.
Rapid Drone Dispatch: Deploying drones to assess the situation and provide aid, such as visual inspection or guiding pickers back to safety.
To evaluate the WatchDog feature, efforts were led by Tampere University (TAU) with close collaboration with Cranfield University (CU), Ingeniarius (ING), and Deep Forestry (DF) during May-August 2024 to design a human-centric test approach. The resulting test approach was made up of two dedicated trials titled “Guide home” and “SOS monitoring”.
Guide home: A feature that guides lost pickers in dense forests back to safety
SOS monitoring: A feature that monitors safe/designated working areas and based on its functional logic, sends a drone out to visually inspect the health of pickers
These WatchDog features were designed to showcase the impact of the system towards enhancing the safety and trust among berry pickers. The first WatchDog field tests were performed across multiple days during the Rovaniemi campaign in the first week of September 2024 with collaboration of TAU, CU, ING, and DF.
All tests utilized the GPS MiniFinder, Drones (Avular Vertex94 and Scout v2), and relevant software modules (e.g., Picker App and Fleet Management System) while pre-test and post-test questionnaires as well as Empatica EmbracePlus smartwatches were used to collect data for psychometric data analysis and help with workload assessments. The preliminary results of the Rovaniemi WatchDog tests are documented in deliverable D1.1 (Human Requirements-Phase 1 + Phase 2) as well as D5.1 Monitoring, navigation and locating validation report Phase 1 Use Cases.
While the WatchDog feature is a development to support the safety of berry pickers, its continued refinement is essential.
Challenges include:
Maintaining a secure connection between the drone and the base station can be challenging due to fog, heavy rain, etc.
Drone battery life directly impacts the coverage area of the WatchDog feature.
WatchDog feature should be capable of communicating with the emergency services efficiently using the standardized message structure.
Obtaining comprehensive insights into the effectiveness of the WatchDog feature due to the limited sample size of test subjects.
Developing more targeted briefings for test subjects to ensure clarity and understanding of the feature and testing process
The field studies have revealed meaningful insights, underscoring the need for continuous testing and adaptation. Additional validation tests are under consideration for 2025, with an expanded participant base and system refinements expected to amplify the gains in safety and trust. Future tests will explore additional automation features, aiming for a fully autonomous SOS response that further minimizes human intervention while maximizing response reliability.