TAIPEI (Taiwan News) — The Industrial Technology Research Institute and the Taiwan Livestock Research Institute have jointly developed a manure-cleaning robot designed for local dairy farms, saving about NT$370,000 (US$11,400) in labor costs annually and reducing daily water use by 13 tons.
The robot features fully automated navigation, allowing it to avoid cattle and obstacles while cleaning waste, then return to its base station for recharging. It can simultaneously scrape and scrub manure while automatically adjusting its cleaning intensity, per CNA.
According to the livestock research institute, farm floors are typically covered with grooved rubber mats to prevent cows from slipping. However, Taiwan’s high-density dairy farms make them challenging to clean, leaving manure trapped in the grooves, which can create slippery conditions and raise the risk of disease.
Imported cleaning machines used by Taiwanese dairy farmers often remove only surface waste and struggle to clean grooves. Their high cost and reliance on foreign maintenance make the new robot a more efficient and affordable cleaning solution.
The robot will be tested at dairy farms in northern and southern Taiwan next year to verify its performance under different climate conditions, after which its functions will be further optimized.
The two institutes have also introduced an electric feeding machine equipped with a dual-motor system that can distribute and push feed while autonomously navigating and recharging. It supports a multiple small-meal feeding model that can boost milk production by up to 30%.
The device can be customized for farms of different sizes. Compared with traditional diesel-powered equipment, it helps reduce air pollution and noise.
Lastly, the two agencies have developed a cow vocal recognition system that can interpret cattle emotions and provide alerts for calving. The system has achieved an 80% accuracy rate in detecting voice during calving and over 70% in emotion recognition. It has been tested at farms in Pingtung and Changhua.
