09/17/2026 | Press release | Distributed by Public on 09/17/2026 13:14
Computer Science and Engineering Professor Wan Du and his students Kang Yang, Yuanlin Yang, Yuning Chen, and Sikai Yang, along with UC San Diego Electrical and Computer Engineering Professor Xinyu Zhang, have received the Best Paper Award at ACM/IEEE SenSys 2026 (International Conference on Embedded Artificial Intelligence and Sensing Systems).
Their paper, "SoilX: Calibration-Free Comprehensive Soil Sensing Through Contrastive Cross-Component Learning," was selected from 505 submitted papers, from whichonly 99 were accepted. Their work was honored as the conference's sole Best Paper Award recipient.
SoilX is a wireless sensing system that measures six key soil properties - moisture, nitrogen, phosphorus, potassium, organic carbon and soil texture - in real time. Existing wireless soil sensors must be recalibrated whenever they are moved to a field of a different soil type. This requires sending soil samples to a laboratory and waiting weeks for results. SoilX eliminates that step with a new machine-learning technique, called Contrastive Cross-Component Learning, that untangles how different soil components interfere with one another's signals, along with a novel pyramid-shaped antenna array that allows sensors to be buried in any orientation.
In laboratory and field experiments across five soil types, SoilX reduced measurement errors by 23.8 to 31.5 percent compared with the best existing system and remained accurate in new fields.
By giving farmers continuous, laboratory-quality information about soil, SoilX could help them irrigate and fertilize exactly when crops need, raising yields while savingwater and reducing the fertilizer runoff that pollutes rivers and lakes. That precision matters in water-stressed agricultural regions such as California's Central Valley, wheresmall errors in sensor readings can translate into wasted water and fertilizers.Because the system is built on low-cost, low-power wireless technology, it could also put advanced soil analysis within reach of small farms that cannot afford frequent laboratory testing.