Speakers

Robot system integrators build systems by incorporating existing sensors. Integrating sensors that are widely available on the market into robotic systems poses no particular difficulty. However, innovative new sensors face significant barriers to adoption in robotic systems due to high costs resulting from low procurement volumes and the lack of a proven track record.

By integrating force-sensing sensors developed by our research team through many years of research into robotic systems, we demonstrate the feasibility of creating robots equipped with skin-like physical sensing capabilities in their legs and body. Specifically, this enables robots to walk on slippery surfaces, stably grasp tools with their hands, and endow robots with force perception.

Furthermore, to reduce manufacturing costs as well as computational and communication costs, we propose an implementation method for locally processing information from a large number of sensors. 

Optical interconnect plays an increasingly important role in data centers and high-performance computing systems. As the I/O bandwidth continue to grow exponentially, traditional electronic switches become the bottleneck of the interconnect network. Optical switching provides direct optical connections between processors with low power consumption, low latency, and unlimited bandwidth. Previously, high radix optical switches can only be made using free space optics. In this talk, I will introduce silicon photonic MEMS (micro-electro-mechanical systems) technology. By integrating MEMS switching elements with silicon photonics, high-radix switches can be integrated on a single chip. In addition to data center networks, high-radix switches also enable a new type of solid state LiDARs. I will discuss our recent work on solid state LiDARs with focal plane switch arrays.

Precision medicine depends on molecular measurements that are sensitive, multiplexed, and actionable on the timescale of clinical decision-making. Yet many genetic and epigenetic assays remain too slow, costly, or infrastructure-dependent to be used routinely at the point of care or in low-resource settings. In this talk, I will present micro/nano-enabled diagnostic systems that convert complex molecular assays into portable, affordable, sample-to-answer workflows.

The core strategy integrates three complementary elements: (i) digital partitioning in microfluidic arrays and droplets to detect rare targets, achieve single-molecule/single-cell sensitivity, and resolve biological heterogeneity; (ii) melt‑coded sensing as an information-rich transducer that enables compact multiplexing with simplified optical hardware; and (iii) magnetofluidic automation using magnetic beads to integrate nucleic acid extraction, washing, and amplification/detection in low-cost cartridges with portable instrumentation. I will highlight applications spanning two time-critical clinical needs: rapid pathogen identification with antimicrobial resistance testing to guide evidence-based therapy, and high-precision DNA methylation analysis for early cancer detection from scarce circulating DNA. Together, these examples illustrate how digitization, melt coding, and cartridge automation can help democratize precision diagnostics by bringing actionable molecular results closer to patients.