Database Reference
In-Depth Information
amounts of soft drugs, are legal under Dutch law, local police have limited means to
stop this criminal expansion. Retailing grams of cannabis and other drugs to thou-
sands of tourists, however, requires a supply-chain of larger quantities.
As the problem had a clear community-transcending character, meetings were
organized between local and national police officers to share knowledge about the
phenomenon and discuss intervention strategies. Participants included criminal in-
vestigators, neighborhood police officers, highway patrol officers, and information
and intelligence specialists. Through these meetings it became clear that Rotterdam
was an important distribution center for heroine and other drugs. Moreover, it ap-
peared that certain groups had specialized in import or retail, while trafficking the
drugs from hidden stashes to retailers was the domain of other groups. While the im-
port of larger quantities of drugs is very irregular and well hidden, and because the
ultimate retail of small amounts of soft drugs is legal under Dutch law, it was rea-
soned that if the police would be able to discriminate between normal traffic and
drug trafficking the criminal chain could potentially be most vulnerable during trans-
port. Given the limited stock and high turnover of coffee shops, trafficking would be
routine. Moreover, if the network of drug-traffickers could be made visible, it could
provide clues about the location of drug-stashes, middlemen, and routes.
Follow-up sessions resulted in the construction of a 'drug-trafficker model'
and profiles consisting of lists of indicators that officers could apply in specific con-
texts. Aided with these profiles several control actions on highways were organized,
stopping and checking hundreds of vehicles. The results were very poor. Just grams
of heroine were found. Clearly, human senses were not well suited to discriminate
drug-trafficking behavior in large traffic flows, while most indicators were only as-
sessable after a vehicle was stopped. As a consequence, the next question was how
the spatial behavior of drug-traffickers could be discriminated from other vehicles 'in
the flow' (with a density of circa 4000 per hour). What data was to be assessed and
analyzed and how could this contribute to the ability of police officers to discrimi-
nate red-handed drug-traffickers from other travellers?
The solution was found in using a real-time complex event processing system,
fed by live data-streams of automatically read license plates, assessed at four strate-
gically chosen points along the route. One of the constructed profiles was aimed at
detecting vehicles that travelled to and from Rotterdam and Maastricht within short
periods of time, a pattern that investigation officers knew to be typical for drug-
traffickers. The profile was further strengthened by combining this information with
a list of license plates of vehicles that frequented coffee shops in Maastricht. This list
could also have been generated automatically, if sufficient sensors would have been
available. For the above profile data needed to be kept in memory for a short period
only (number of hours). To reduce the impact of privacy invasion, reads of license
plates that within this period did not score on the profile were automatically re-
moved. Moreover, as the profile was based on time-spatial behavior only, the profile
was discrimination-free.
At the beginning of the operation more seasoned policemen were sceptical
about the idea that technology could complement their sentience. Initially, thousands
Search WWH ::
Custom Search