Novel encoder for ambient data compression applied to microcontrollers in agricultural robots

Agricultural robots are flexible to obtain ambient information across large areas of farmland. However, it needs face two major challenges: data compression and filtering noise. To address these challenges, an encoder for compression, named Tiny-Encoder, was presented compress filter raw information, which can be applied agricultural robots. Tiny-Encoder is based on the operation convolutions pooling, has a small number layers filters. With aim evaluating performance different three types (including temperature, humidity, light) were selected show compressing In task data, obtained higher accuracy (less than maximum error sensors ±0.5°C or ±3.5% RH) more appropriate size (the largest 205 KB) other auto-encoders convolutional operations with compressed features 20, 60, 200 features). As noise, comparable conventional approaches median filtering, Gaussian Savitzky-Golay filtering). kernel (i.e., 5), best among four approaches: coefficients variation 5) 8.6189% (temperature), 10.2684% (humidity), 57.3576% (light), respectively. Overall, used microcontrollers in acquisition Keywords: robot, embedded machine learning methods DOI: 10.25165/j.ijabe.20221504.6911 Citation: Wang Y T, Li M Z, Ji R H, J, Zhang Y, Zheng L H. Novel Int J Agric & Biol Eng, 2022; 15(4): 197–204.