Sorry, you need to enable JavaScript to visit this website.

The Energy-Smart Farm: Distributed Intelligence Networks for Highly Variable and Resource Constrained Crop Production Environments

The Energy-Smart Farm: Distributed Intelligence Networks for Highly Variable and Resource Constrained Crop Production Environments
February 13-14, 2018
Phoenix, AZ

ARPA-E hosted a workshop entitled “The Energy-Smart Farm: Distributed Intelligence Networks for Highly Variable and Resource Constrained Crop Production Environments,” on February 13-14, 2018 in Phoenix, AZ. 

Among the great strategic assets possessed by the United States of America are its vast supplies of arable forest, crop and rangeland capable of providing abundant energy and food feedstocks for domestic consumption and international export. However, a 60-100% yield gap exists between best management crop production practices and state averages, largely due to variation in field environment and suboptimal field management decisions. Farms are getting bigger, workforces smaller, and while digital agriculture holds immense potential for much-needed gains in energy and resource efficiency, today’s data tools are limited in power, resolution and cost.

Meeting projected demands for fuel, food and fiber in the 21st century requires a transformation that will sustainably double crop productivity in the face of competition for arable land and fresh water and increased exposure to climatic shocks. This workshop will convene experts in the biological, physical and computational sciences to explore emerging opportunities in advanced sensing systems for energy-smart agriculture. The overall challenge is to monitor the physical, environmental and biological conditions that limit growth at high spatial and temporal resolution throughout production cycle and to identify interventions that relieve those constraints.  Should this become a funding opportunity, teams will be asked to develop innovative technology suites and decision support tools that maximize sustainable economic returns by conserving resources and energy, increasing yield, and gathering critical data for life cycle analysis and participation in environmental service markets.

Technologies of interest are intended to support the development of fully-integrated, low-cost sensor networks with connectivity standards for user-friendly, turn-key solutions in highly-variable, resource-constrained environments. Specific interests include: 

• Sensors: a range of biological, chemical, and mechanical sensors (e.g. CNT, photonic crystal, etc.) that have potential to be downscaled with minimal precision sacrifice and can withstand a wide range of conditions.
• Low-cost, small-scale computing “at the edge”: High-resolution distributed data generation, local computation and secure transmission of information in remote locations.
• Wireless sensor networks: Overcoming technical barriers in secure, reliable, and energy-efficient connectivity and coordination.
• Multiscale, multimodal sensor deployment and integration: Data fusion techniques to register remote-sensor, point-sensor, genomic, weather, and other data for broad-acre usage.
• Analytics and Machine Learning: Algorithms to optimize yields in practice; user-friendly interfaces; permissioned management and secure data transmission on the farm and through product delivery to the supply chain.

Target outcomes include: 

• The identification of specific technical barriers to the development and adoption of the aforementioned tools at acceptable cost to growers.
 Realistic timeframes and technical metrics for successful prototypes.
• New professional relationships among disparate technical communities in the biological, chemical, engineering, and computational sciences, which could form the basis for teaming opportunities.

 Tuesday, February 13 

9:30 am Registration and coffee

9:00 am One-on-One Meetings

12:30 pm Dr. Pat McGrath
Welcome and Introduction to ARPA-E

12:45 pm Dr. Joe Cornelius
Introductory Presentation

1:15 pm Table by Table Attendee Introductions

2:00 pm Dr. Cristine Morgan
Sensing for Yield and Energy-Smart Farming- Soil Management

2:15 pm Dr. Cara Gibson
Sensing for Yield and Energy-Smart Farming – Pest Management

2:30 pm Dr. John Antle
Agricultural Systems Science

2:50 pm Dr. Brian Anthony
SENSE.nano: Lessons from Medicine & Industrial IoT

3:15 pm Dr. Troy Olsson
N-ZERO: Low-Power Sensing

3:40 pm Dr. Frank Libsch & Dr. Robin Lougee
Micro-Sensor Platforms and Machine Learning at IBM

4:00 pm Parker Liautaud
Breakout Overview

4:15 pm Breakout Session 1

Group 1- Bioenergy Crop Production Priorities
Group 2- Abiotic Sensors & Platforms
Group 3 - Biotic Sensors & Platforms
Group 4 - Decision Support Analytics

Wednesday, February 14th 

8:30 am Dr. Michael Wang
Bio-Product Life Cycle Analysis

9:00 am Dr. Kevin Dooley
Supply Chain Alignment of Farm-Level Metrics

9:15 am Raja Ramachandran
Transforming Agricultural Supply Chains with Verifiable Transparency

9:50 am Dr. Ranveer Chandra
FarmBeats: End-to-End Design of IoT for Agriculture

10:20 am Parker Liautaud
Breakout 2 Overview

10:30 am Breakout Session 2
Group 1- Straw Model System Design
Group 2 - Straw Model System Design
Group 3 - Straw Model System Design
Group 4 - Straw Model System Design

12:00 pm Workshop Concludes

Breakout 1 The aim of this breakout is to evaluate the components of an integrated platform for on-farm decision support. Groups will focus on agricultural production priorities, sensor and platforms for data collection, and analytics for rapid, farm-relevant outputs. Attendees will be asked to prioritize research thrusts and evaluate the potential impact that this approach may have on bioenergy production.

Breakout 2 Based on the input from Breakout 1, the aim of this breakout is to build out a notional in-field system for decision support. Attendees will be asked to design a straw model system, outline the required sensors and necessary metrics for those sensors, and evaluate the impact that these novel datasets will have on the farm. Groups will also be asked to consider the implication of these novel datasets in the long term,