• We need to let the WiSAR community see/try our stuff and provide direct feedback. In order to achieve this, we have to:
  • Talk to them more. Instead of always going through Ron, go meet them face to face and tell them what we are doing. Show them presentations, videos, planes. Discuss with them directly to see how they see our stuff could fit in.
  • Instead of inviting them to our field trials, ask them when they would like to see our demo. Get the decision makers involved. We'll work around their schedule.
  • As Dr. Goodrich mentioned, we can go to their training ground to demo. And we don't have to demo a real full search scenario. We can just demo the technology capabilities we have.
• We need to impress them (demonstrate how easy it is and how powerful it is):
  • Easy setup (software, hardware, configuration)
    • Software: this is actually not very easy right now. I think we need to either provide them laptops already with things installed and configured correctly, or we need to create an installation package for our stuff with all or most of the configuration done by default values.
    • Hardware: this is not very easy right now either. We need to show a list of the required components, an easy-to-understand diagram showing how the components are connected and how they work together, and an easy-to-understand instruction (checklist) sheet for setting up the hardware. (We actually need this kind of stuff ourselves at BYU so when students graduate, we still have good documentations)
  • Easy hardware operation (autonomy: HAG, auto Launch, auto landing, etc.)
  • Easy software operation (e.g. intuitive interface, path planning, etc.)
• We need to attract public attention:
  • News media coverage: I often read about UAV research done at other universities. I think we have done a lot more and are doing better, but I haven't really read anything about UAV research done at BYU in the news. The WiSAR community probably would also like more coverage of their work in the news (from a technology improvement sense, not incidents coverage).
  • YouTube videos: MAGGIC lab has posted multiple videos (although the best one, the big compilation, had its audio track disabled by YouTube for copyright issues). A picture is worth 1000 words, and a video is worth 30000 words per second.
  • Web presence: Our wiki web site is restricted to login users only. Is there a way we can make it viewable to the public? We can also have a specific project site where we can post research progress and lots of eye tracking pictures and videos. If we want our stuff to get out, to have real impact, we need to package our “products”. I'd also like to see the current wiki web site organized better. Maybe we can assign a student to be in charge of organizing things on the wiki. (Same goes with HCMI wiki and TiLAR wiki)

Comments by MG. Good comments Lanny. Thank you. I'll respond to the part about attracting public attention. We have decided not to do this until after we have built a relationship with the Utah County Search and Rescue Community. It would be a mistake, in my opinion, to talk to the press before we have this relationship on solid ground. This is the reason that the wiki is restricted to login users only; it is not the intended destination for the site, but just a temporary measure until we build the relationship. In just a minute, I'll post a plan for building the relationship with the WiSAR community.

• Use multiple UAVs: Mike G: Why?
• Support the entire operation (where UAV is only a small component). Develop interface to track search progress, annotate scenarios, collect geo-referenced information from ground searchers and display at central command on maps (Google earth), etc. Mike G: Agreed. Mike Roscheck has just completed the lit review on this and he found several GIS solutions for doing this. He is now working to fill in the gaps in this work. Interestingly, modeling how missing persons move as a function of terrain is a big hole in the current literature. Quoting from “Lost Person Behavior” by R. J. Koester, “A user friendly ground-based computer model that automatically generates a [probability of area] map … does not currently exist.
• Better support the ground searchers (ground searcher uploading images, improve situation awareness by getting overall summary info through portable devices, watch video on demand, IC pushes images/videos to specific ground searchers, etc.) Mike G: Agreed. Mike Roscheck is working on this. A big limitation right now is communications. How do we create a cheap, reliable, low-power network in wilderness settings?
• Could work with other search/rescue domains (USAR, Forest Fire, Flood, Earthquake, Typhoon, etc.)
• Could work with other non-search/rescue domains (if it's easy to quickly adapt our technology) such as surveillance, news coverage, inspection etc.)
• Present our technologies to all kinds of people and let them think about how our technology could benefit whatever they do (except seek and destroy).

• More intelligence (autonomy) (Let's put it this way, if the UAV flies all by itself and intelligently collect and analyze information, then it can suggest things for the search and rescuers to look at. In the worst case, the performance of the search and rescuers will be no worse. But it's possible that with the additional suggested information, the performance can only get better.)
  • Autonomy detection in video
  • Path planning based on autonomy detected
  • Better auto take off and auto landing (e.g. a pre-defined landing spot, so the UAV will come back and land by itself – so it can be safely ignored.)
• Interface ideas
  • Video navigation techniques
  • Flight control techniques
  • Path planning techniques (sliding autonomy)
• Infrastructure (framework) for WiSAR operations
  • How to integrate all components together
  • How to best support the operation as a whole