Industry

Industry Partnerships

The lab partners with industry across sponsored research, consulting, joint proposals, and licensing — with a track record spanning ARPA-E, DOE, and NSF-funded programs and alumni in leadership roles at Google, Apple, Amazon, Nvidia, KLA-Tencor, Cymer, Enphase, NREL, Tesla, Skyworks, GE Research, and IBM Research Zurich.

Alumni in industry

31 lab alumni currently in industry roles. The full list is on the Alumni page.

Past industry partners & sponsors

Direct engagements and equipment grants documented in publications and grant records include:

  • IBM Zurich Research Labs — probe-based data storage (multiple grants, 2006–2009)
  • Asylum Research / Oxford Instruments — equipment and product-related work
  • Google — topology estimation and control of electrical grids (research gift)
  • Dynapower — ARPA-E NODES co-PI (2015–2019)
  • NREL — Rapidly Viable Sustained Grid
  • Digital Instruments — AFM equipment grant (1998)

Programs: NSF (multiple awards on multi-objective control, structured control, less-conservative stability, and information-theoretic limits) · ARPA-E (NODES, Rapidly Viable Sustained Grid) · DOE · NIH · Muscular Dystrophy Association.

What we do for companies

Capabilities the lab has delivered on in prior sponsored, consulting, or joint-proposal work, organized by research area.

Energy

  • Distributed control of grid-forming (GFM) and grid-following (GFL) inverters
  • Cyber-secure microgrid operation resilient to false-data injection
  • Non-intrusive load monitoring at residential and commercial scale
  • Distributed apportioning of energy resources for ancillary services
  • Power hardware-in-the-loop validation of grid-edge control algorithms
  • Reinforcement learning for electricity market vulnerability analysis

Nanoscience

  • High-bandwidth nanopositioning system design
  • Real-time quantitative measurement of elasticity and dissipation at the nanoscale
  • Custom AFM cantilever probe design and characterization
  • Transient Force AFM (invented in this lab)
  • Communication-theoretic modeling and detector design for probe-based data storage
  • Feedback-enhanced identification of optical tweezers

Single-Molecule Biophysics

  • Single-molecule force spectroscopy of proteins
  • Statistical analysis and event detection in single-molecule time series
  • Semi-analytical modeling of intracellular cargo transport
  • Physics-augmented machine learning for force-spectroscopy data

Foundations

  • System identification of dynamical networks from time-series data
  • Network topology reconstruction with unobserved nodes
  • Causal discovery from passive and partial observations
  • Multi-objective and structured robust controller synthesis
  • Distributed optimization and consensus algorithms

Machine Learning

  • Physics-augmented deep learning for scientific and engineering data
  • Reinforcement learning for security-critical control
  • Sample-complexity-optimal learning of dynamical DAGs
  • Data-driven system identification with provable guarantees

How we engage

Sponsored research

Multi-year research programs administered through the UMN Office of Sponsored Projects. Suitable for open-ended problems where the deliverable is a body of research.

Consulting

Individual consulting engagements under UMN's faculty consulting policy. Best for focused, defined-scope technical questions.

Joint proposals

Partnering on SBIR/STTR, ARPA-E, DOE, and NSF Industry-University Cooperative Research proposals.

Contract measurement

Use of lab facilities (AFM, optical tweezers, PHIL testbed) for defined measurements. See Facilities.

Licensing

IP developed in the lab is available for licensing through UMN Technology Commercialization.

Advisory

Technical advisory relationships with early-stage companies and program advisory roles for larger organizations.

Selected case studies

  • ARPA-E NODES — grid architecture for renewable integration

    Problem: Widespread renewable generation stresses grid reliability and power quality; existing centralized control cannot scale to millions of grid-edge devices.

    Approach: A distributed control framework with plug-and-play architecture supporting both centralized cloud-based and peer-to-peer coordination, applying nonlinear and robust control theory to self-organizing power systems.

    Outcome: Framework tested with 100+ physical devices (photovoltaics, battery inverters, home appliances) at the lab's PHIL testbed. Multiple publications; alumni (Blake Lundstrom, Govind Saraswat, Sourav Patel) now applying the framework at Enphase, NREL, and Google's data-center carbon-neutrality efforts.

  • IBM Zurich — dynamic-mode probe-based data storage

    Problem: Probe-based data storage at high densities requires channel models and detector designs that account for the physics of tip-media interaction and integrated thermal sensing.

    Approach: Developed communication-theoretic models for probe-based data storage, ML sequence detectors tailored to the observed channel, and identification of electrostatically-actuated cantilevers with integrated thermal sensors.

    Outcome: Multiple joint publications with IBM Zurich; two grants sponsored the work (2006–2009).

  • Real-time quantitative material properties in AFM

    Problem: Existing AFM modes could not simultaneously image topography and quantitatively determine material properties (elasticity, dissipation) in real time for soft matter.

    Approach: A new real-time methodology combining systems-theoretic modeling of the tip-sample interaction with simultaneous parameter estimation.

    Outcome: First reported real-time quantitative material property measurement in dynamic-mode AFM. Method is compatible with existing AFM hardware and has been used across polymer and biological samples.

Get in touch

For industry inquiries, email murtis@umn.edu with a subject line beginning "Industry inquiry:" and a brief description of the problem, timeline, and preferred engagement mode.