We herein report on our recent investigation on the development of flexible control of micro (fluidic) system by the application of machine learning (reinforcement learning). Classical peristaltic micropump is employed as platform of the studies and Q-learning algorism, which is also classical algorism of reinforcement learning, is applied to the system. The acquiring of optimal micropumping behavior and manipulation of microbead in microchannel are demonstrated on the platform. The acquired micropumping sequence realize higher flow rate than typical sequences proposed in earlier studies. It is understood that the unique characteristics of the system are considered to acquire the sequence. The efficient micromanipulation of the microbead is also demonstrated on the same platform, even the microdevice is originally designed for the micropumping. Therefore, it could be concluded that the application of the reinforcement learning to a microsystem could be effective to extend the versatility by bringing out the potential of the system.

A wrist-watch type biosensing system for real-time sweat lactic acid (LA) monitoring was fabricated and tested. The system consists of a microfluidic LA biosensor, a carrier-flow supplying device, and a sampling device for transporting whole secretions at the skin surface to the biosensor by continuous flow of phosphate-buffered saline (PBS). The perfusion test of PBS onto the skin surface by using the sampling device resulted in less than 1% leakage frequency at the collection site. The carrier-flow supplying device was able to feed PBS regardless of the body angle. When the subject wearing the device performed resistance training (lateral raise, 20 repetitions, 12.5 kg dumbbells), the LA secretion increased from 2 µg/cm²/min to 16 µg/cm²/min in response to the exercise, and then recovered to the baseline. The setup time for the measurement was less than 60 seconds. This indicates that the wrist-watch device can readily measure LA dynamics on the skin surface under conditions of the subject.

Droplet microfluidics is a powerful tool for high-throughput experimentation, and droplet coalescence is necessary for mixing and chemical reactions. Droplet merging and polymerization will release the limits on particle synthesis and widen the technical potential of droplet microfluidics. Previously, a focused laser beam has been used to induce the coalescence of droplets. This paper reports the parallel photothermal coalescence of biocompatible photocurable polyethylene glycol diacrylate (PEGDA) droplets using a 2D image for developing a more efficient coalescence process. PEGDA droplets with diameters of 18 µm to 50 µm were generated in a microfluidic flow-focusing device and stored in a microchannel. A 2D image of violet light induced the parallel coalescence of PEGDA droplets with diameters of around 30 µm. When continuous-phase oil was replaced with nitrogen, PEGDA droplets were photopolymerized.