As part of the Sensors, Machine vision, Automation and Robotics for Transforming Plants (SMART Plant) project, ORNL researchers demonstrated how integrated automation, robotics, imaging, sensors, and real-time edge computing capabilities can accelerate scientific discoveries in field experiments, providing insights into strategies for enhancing aboveground plant productivity, belowground soil carbon storage, and monitoring ecosystem responses.
Transforming the current pace and scale of measurements belowground is critical to address knowledge gaps about the how spatial and temporal variability in soils affects carbon and nutrient cycling. To meet the grand challenge of sequestering atmospheric carbon dioxide in soils at landscape-scales, more data are needed.
Data gathered during the project can inform predictive models on responses of managed and natural ecosystems, and advance research and development of bioenergy crops with higher carbon storage capacity above and belowground.
Demonstrating a new, autonomous approach to field science
As part of this SMART Plant F-series project, researchers designed and built a small, all-terrain robot that can dig and collect soil samples. The robot sampled at a depth and volume that scientists study to understand how carbon, nutrients, and other factors vary belowground in different locations. The site, located on Oak Ridge reservation, was the first SMART bioenergy site of its kind to collect data from in-ground, on the ground, mobile, and aerial sensors. Onboard sensors sent data about the soil samples to a computing station at the field in real-time.
Scientists used the ground-truthing data the system collected and compared it to samples secured through more traditional, manual methods. The field was planted with 200 variants of poplar, a fast-growing bioenergy feedstock, and researchers evaluated soil, root, and leaf samples collected from the field site.
The team developed the necessary cloud communication and data handling workflows to enable the edge computing station to collect the data from these aerial, on-ground, and belowground sensors. Under the ORNL INTERSECT initiative, Udaya Kalluri and team developed computational microservices for real-time automated assessment of the sample data and autonomous steering of experiments.
This proof-of-principle project demonstrated how automation can speed scientific discoveries in the field. It is the first step in producing a lab-to-field plant performance prediction capability to accelerate development of better crops in support of a thriving bioeconomy.
-
Pre-planting scoping of field site by cross-cut research team of robotics and plant researchers.
-
Researchers used a special lidar camera to gather ground-level information at the poplar site.
-
Researchers used a drone to collect aerial data on the poplar field.
-
Pre-planting scoping of field site by cross-cut research team of robotics and plant researchers.
-
Researchers used a special lidar camera to gather ground-level information at the poplar site.
-
Researchers used a drone to collect aerial data on the poplar field.
