Introduction

SARBayes is about how to generate accurate probability maps, and use them quickly, automatically, and well to find missing people on land more quickly. It is about applying Bayesian statistics more broadly to SAR. We develop models and algorithms to support missing-persons searches on land, using data that we and others have collected about missing-person behavior and search operations.

Some of our more notable projects include:

• collecting data on Australian searches into an online database
• making and testing Bayesian network models to predict lost-person behavior
• developing a general-purpose library for optimal resource allocation

SARBayes is a research project organized by Charles Twardy. It began in the Reasoning Under Uncertainty Group, a part of the Monash Data Mining Centre) at Monash University in Melbourne, Australia, with the cooperation of the Victorian Police Search & Rescue Squad, and VicWalk's Bushwalkers' Search & Rescue. The project has received invaluable assistance from other individuals and organizations, especially Bob Koester, Jack Frost, and Alan Washburn.

The unifying theme of SARBayes is Bayesian inference: search is a classic case of Reasoning Under Uncertainty and the core of the problem is generating and maintaining a probability map for the current location of the lost person. Search theory and optimal resource allocation presume that such a map exists, but there has been up until now no good way to make one for land search.

The database of lost-person incidents is one approach to generating such a probability map — by matching current case data to a statistical profile generated from past incidents. We use Bayesian networks to turn the bare statistics into dynamic models, and hopefully improve on the very rough statistical profiles now used. There are other approaches besides case-history, such as modified random walk simulations, and of course, expert scenario analysis at the scene. An ideal system would combine all of these approaches. And Bayesian statistics provide a clear and principled way to do that too.

Our goal is to be of real use to the SAR community. We expect our models to beat the current state of the art, and we expect our software to make it possible to use those models on real searches. And at a minimum, we provide an ongoing database of Australian cases, from which others can develop models and methods.

Comments are welcome (remove the underscores): c_t_w_a_r_d_y at sarbayes.org

News

18 Nov 2007: Blog: The navbar now links to the blog. Someday, I'll figure out how to make the blog handle this news bar too.

22 Feb 2007: CVSDude for SORAL: The SORAL source is going online with its own trac site on CVSDude. Bug-tracking and Wiki are up. Code to arrive soon.

25 Nov 2006: Hosting Revolution: Several changes. We are moving to Hosting Revolution, re-energizing SORAL development, and putting up some other reports: the draft reports for Washington State and New Zealand (2004-2005 data).

7 July 2006: Data & Programs: We've posted the Aussie data (with identifying info removed) and the Python programs. See Downloads.

29 June 2006: NATSAR Report: At last, we release the final version of "Missing Person Behaviour: An Australian Study", almost 3 years after the draft report. 550 cases, 458 qualifying for analysis. 64 pages, including a background chapter reviewing previous studies. Available at Downloads. Goes to NATSAR in November.

31 March 2006: Following a permanent server crash at Monash, we've moved to Earthlink. We apologize for the month of downtime. The new site should be more reliable. Please bear with us as we adapt to the new host configuration. The static pages should work, but the database, Wiki, CVS, and CVSTrac will take some fixing.

Early 2005: Apologies. Work has been slow in 2004. With luck we'll finish the final draft of last year's report to NATSAR soon. Some new data, but not as much as we'd hoped.

25 Nov 2004: Updated links page. Now has a navbar, and some links to software modelling exposure. Slight reorganization, some new news on SARMaster, etc.

12 Nov 2003: Preliminary Report "Missing Person Behaviour: An Australian Study" goes to National SAR Council (NATSAR)! 348 entries, 288 cases. 25 pages. Some clear trends, and a clear call for NSW, ACT, and NT to supply their data. We're not releasing because of missing data, but if you would like to review and comment on the draft, email Charles.

9 May 2003: SORAL 2.0 Beta_2 released GPL! See the Downloads pages. Passes all tests, all major design changes made, with new User's Guide and Developer's Guide!