CPS: Medium: Leveraging Honey Bees as Bio-Cyber Physical Systems
pdf
The goal of this project is to leverage and improve upon the capabilities of honey bees as agricultural pollinators by incorporating them into Bio-Cyber Physical systems. Rapid advances are needed to aid a dwindling agricultural workforce, increase crop yield to sustain the growing population, and provide targeted crop care to limit the need for broad pesticide treatments. These challenges may well be addressed by autonomous mobile robots and sensor networks; unfortunately, agricultural landscapes represent vast, complex, and dynamic environments that complicate long term operation. In contrast, social insects are capable of robust sustained operation in unpredictable environments far beyond what is possible with state-of-the-art artificial systems. Colonies of honey bees are of particular interest, because they are the premiere agricultural pollinator bringing in over $150 billion annually. A colony causes pollination by dispatching tens of thousands of scouts and foragers to survey and sample kilometer-wide areas around their hive. Thus, the colony as a whole accumulates vast information about bust and bloom in the local agricultural landscape -- information that would be very helpful to farmers and beekeepers.
To harness the capabilities of a bee colony while still providing control and sensing, the proposed work specifically involves 1) novel submillimeter flight recorders with visual scene capture and analysis, thermal and mechanical sensors, a clock, storage, processing, photovoltaic chargers and short range communications; 2) algorithms and models to estimate foraging maps, relying on bee motion models and feature extraction, merging probability density functions of observed landmarks from thousands of flights; and 3) feedback control via a bee-mimicking shaker device to recruit foragers, in turn eliciting data collection and pollination, e.g. during brief spouts of bloom that would otherwise go unnoticed by the colony. This research represents a transformative step towards a new frontier in Bio-Cyber Physical Systems, improving upon the abilities of social insects to sense and interact with the physical world, while providing data acquisition and control on par with explicitly engineered systems.
Over the past two months, existing angle-sensitive pixels out of Prof. Molnar’s lab have been interfaced with a stand-alone microprocessor board. This board is currently being installed on a drone which will perform bee-mimicking flights to test the viability of automatically computed flight probability maps. The drone flight patterns are being verified against real honey bee flights which were measured by automated visual tracking on video recordings of honey bee foragers trained to a feeder station. The drone-tests will inform the design of the first iteration of micro-scale flight recorder chips to be produced in the beginning of the following season. To further inform models of what sites foragers advertise in the hive, infrared videos of observation hives have been recorded. The next step is to develop automated visual trackers to capture data on where and what sites are advertised, to be compared against data on where honey bees actually scout. Finally, an initial version of the bee-mimicking shaker device has been tested in the field, and will be automated over winter to be tested in a full-scale experiment in the summer of 2018.
Submitted by Kirstin Petersen
on
The goal of this project is to leverage and improve upon the capabilities of honey bees as agricultural pollinators by incorporating them into Bio-Cyber Physical systems. Rapid advances are needed to aid a dwindling agricultural workforce, increase crop yield to sustain the growing population, and provide targeted crop care to limit the need for broad pesticide treatments. These challenges may well be addressed by autonomous mobile robots and sensor networks; unfortunately, agricultural landscapes represent vast, complex, and dynamic environments that complicate long term operation. In contrast, social insects are capable of robust sustained operation in unpredictable environments far beyond what is possible with state-of-the-art artificial systems. Colonies of honey bees are of particular interest, because they are the premiere agricultural pollinator bringing in over $150 billion annually. A colony causes pollination by dispatching tens of thousands of scouts and foragers to survey and sample kilometer-wide areas around their hive. Thus, the colony as a whole accumulates vast information about bust and bloom in the local agricultural landscape -- information that would be very helpful to farmers and beekeepers.
To harness the capabilities of a bee colony while still providing control and sensing, the proposed work specifically involves 1) novel submillimeter flight recorders with visual scene capture and analysis, thermal and mechanical sensors, a clock, storage, processing, photovoltaic chargers and short range communications; 2) algorithms and models to estimate foraging maps, relying on bee motion models and feature extraction, merging probability density functions of observed landmarks from thousands of flights; and 3) feedback control via a bee-mimicking shaker device to recruit foragers, in turn eliciting data collection and pollination, e.g. during brief spouts of bloom that would otherwise go unnoticed by the colony. This research represents a transformative step towards a new frontier in Bio-Cyber Physical Systems, improving upon the abilities of social insects to sense and interact with the physical world, while providing data acquisition and control on par with explicitly engineered systems.
Over the past two months, existing angle-sensitive pixels out of Prof. Molnar’s lab have been interfaced with a stand-alone microprocessor board. This board is currently being installed on a drone which will perform bee-mimicking flights to test the viability of automatically computed flight probability maps. The drone flight patterns are being verified against real honey bee flights which were measured by automated visual tracking on video recordings of honey bee foragers trained to a feeder station. The drone-tests will inform the design of the first iteration of micro-scale flight recorder chips to be produced in the beginning of the following season. To further inform models of what sites foragers advertise in the hive, infrared videos of observation hives have been recorded. The next step is to develop automated visual trackers to capture data on where and what sites are advertised, to be compared against data on where honey bees actually scout. Finally, an initial version of the bee-mimicking shaker device has been tested in the field, and will be automated over winter to be tested in a full-scale experiment in the summer of 2018.