Balancing Cost and Reliability in Autonomous Agricultural Robotics: A Case Study in Horticulture

<p>Despite immense potential, high costs limit agricultural robot adoption, particularly for smaller farms. This research explores more affordable robots, sensors, and navigation solutions. We compare two robots of differing costs, evaluating their performance to understand the cost-effectiveness trade-off in agricultural robotics. A novel deep learning-based visual navigation system for low-cost cameras is featured, offering an affordable alternative to traditional high-precision methods (e.g., LiDAR, GNSS) in polytunnels. This comparison benchmarks our developed autonomy stack's performance, not the commercial viability of the robots themselves. Though our low-cost system shows a higher intervention rate than the commercial prototype, requiring more human supervision, its significantly lower cost enables deployment of multiple robots. This approach can achieve the same fruit transportation task as a single high-cost robot with less upfront investment. This cost-reliability trade-off opens new avenues for farms with limited budgets to deploy agricultural robots.</p>