PI: Greg Lowry
Co-PI(s): Aaron Johnson
University: Carnegie Mellon University
Industry partner: Geosyntec Consultants
Characterizing the types and extent of soil contamination over vast areas such as old industrial sites or regions affected by natural disasters like massive flooding is time and labor intensive and expensive. Adapting appropriate sensing technologies onto robotic platforms, coupled with novel artificial intelligence algorithms for smart and sampling, have the potential to automate soil contamination characterization, making it faster and lower cost. However, this requires automating sampling and analysis methods for soil contaminants and development of smart exploration algorithms that enable autonomy. The proposed PITA project converges expertise in robotics and environmental engineering at CMU with expertise in remediation engineering at a PA environmental consulting firm (Geosyntec) to develop robots that can autonomously characterize and map the distribution of heavy metal contaminants in soils at Brownfield sites in Western PA being considered for redevelopment. Lowering the cost of characterization will enable more redevelopment of these sites. The autonomous sampling methods developed in the study are extensible to additional challenges such as rapidly delineating risks of contaminated surface soils after natural disasters or determining the success of a remediation action.
The specific goals of this PITA project are to
- Determine the soil properties and environmental variables that affect the sensitivity and accuracy of measurements of heavy metal contaminants in soils using automated X-ray Fluorescence (XRF) measurements
- Develop exploration algorithms designed for specific site characterization goals that can make the process autonomous
- Demonstrate the utility and cost savings of the approach at real contaminated sites to create market demand