ºÃÉ«ÏÈÉúTV

The 125 Ideas That Shape the Future archive highlights featured technologies previously showcased by CTTEC as part of ºÃÉ«ÏÈÉúTV’s 125th anniversary year.

These spotlights celebrate CMU research and translational innovation across a range of fields, from public safety to artificial intelligence.

Archive Spotlight

Traffic Jam: Technology with Purpose

Traffic Jam is a platform developed at the ºÃÉ«ÏÈÉúTV Robotics Institute and commercialized by Marinus Analytics.

From Graduate Research to Global Impact

What began as a graduate research project at ºÃÉ«ÏÈÉúTV’s Robotics Institute has grown into a critical public-safety technology used by law-enforcement agencies and nonprofit organizations across North America, South America, and Europe.

Early support from the National Science Foundation helped advance the research, demonstrating how federally funded university work can translate into scalable tools that protect vulnerable populations.

Built for Real-World Investigations

Traffic Jam is designed to support victim-centered, trauma-informed investigations—helping professionals focus time and resources where they matter most.

• Identify indicators of human trafficking
• Detect and disrupt organized criminal networks
• Locate missing persons
• Analyze and map trafficking activity to inform resource deployment

The platform has served more than 5,000 law-enforcement and nonprofit users, analyzed more than one billion online records, and helped reduce investigative timelines from years to months.

Archive Spotlight

Accelerating Vision Transformers with Adaptive Patch Sizes

The first technology highlight in the 125 Ideas series corresponds to the publicly available research paper, “Accelerating Vision Transformers with Adaptive Patch Sizes”, authored by PhD student Rohan Choudhury under the co-advisement of Professors Kris M. Kitani and László A. Jeni in the Robotics Institute.

The work emerges from the combined research environment of the Cognitive Assistance Laboratory and the Computational Behavior (CUBE) Lab, supporting high-performance, human-aware AI systems capable of interpreting and responding to complex real-world environments.

Creators

• Rohan Choudhury — Graduate Student, Robotics Institute
• Jung Eun Kim — Visitor, Robotics Institute
• Jinhyung (David) Park — Doctoral Student, Robotics Institute
• Eunho Yang — KAIST
• László A. Jeni — Robotics Institute
• Kris M. Kitani — Robotics Institute