Postal operators, for example, the US Postal Service (USPS), or their business partners
could offer a novel category of data collection services arising from the ubiquity and
route structure of the postal delivery fleet. This chapter, for the first time, proposes that
mobile sensors mounted on postal trucks could collect and aggregate a variety of important
data as a byproduct of postal delivery, taking advantage of efficiencies of scope and
scale. The data collected might include, among others, air pollution levels, weather data,
sensing of chemical and biological agents, and areas of weak cell phone service. If the
market challenges could be addressed, these services could provide substantial public

Section 2 provides a description of the technologies involved and the basic rationale
for integrating sensor network services with postal operations. Section 3 describes potential
applications. Section 4 discusses market aspects and mechanisms for developing
functioning markets. Section 5 provides conclusions.

Sensors mounted on postal trucks could collect and upload important local data as a
byproduct of postal delivery. USPS or its business partners could lease out space on
selected postal trucks to permit installation of smart sensors with radio linlcs to collect
and transmit various types of data, and record the location and time of each piece of
data. Postal routes are tailor-made for a sensor network because postal delivery routes
reflect locations of human activity and the trucks traverse those routes daily. The existing
set of routes closely resembles the structure of an efficient ground-based mobile sensor
network designed from scratch.

The concept would not add appreciable labor hours, route diversions or changes in
procedure. It would not add tasks to the drivers' workload that would detract from mail
delivery, although drivers may be able to provide feedback about proper functioning
of the sensor system. While sensors at fixed locations are con,fined by functional and
geographic limitations, a mobile sensor network using postal trucks can provide tightly
interlaced, overlapping fine-grained coverage across a broad area. It constitutes a virtual
(cloud' of sensors spread throughout a region. Individual sellsor points move around, but
the aggregate cloud remains deployed and well dispersed much of the day, most days of
the week. A network of mobile sensors can reach nearly everywhere in the country, yet
can also achieve a high degree of geographic specificity and selectivity. It permits targeting
of routes near a specific subset of industrial facilities, specific types of neighborhoods
or lists of critical infrastructure points, depending on customer needs. Factories producing
or using a particular chemical may require monitoring nationally, and to do this a
small subset of neighboring postal routes can be selected.

Measurements can be done nationally, regionally, locally and selectively down to an
individual route. The systeln allows collection of finely or coarsely spaced data. The
sensing area can cover a wide swath of the nation or just a single metropolitan area, or
a location conforming to a set of specific target criteria, even locations cutting across
sections of specific postal routes.

The rate of deploynlent is also flexible - sensor packages can be deployed simultaneously
across a set of trucks for rapid data collection, or else in stages. Thus, the number
of sensor packages is adaptable - tIle number of trucks required for a particular task
can be reduced by gathering data on a progressive basis. Once a route is run the requisite
nunlber of times and adequate data are obtained, the sensor packages can be moved to
another set of routes. This provides for great utility in sensor deployment and extraction
of maximum value from their use.

Flexibility of the concept is enhanced by availability of several data acquisition
options. Mechallisms may include accumulating or averaging data over time to establish
a baseline or to locate irregular phenomena, adapting the deploynlent and use of sensors
based on the data received, and accommodating multiple sensors on tIle same vehicle
platform. Data can be recorded and provided continuously, intermittelltly or on predetermined
trigger points, or once daily. Data can also be used to trigger alerts/alarms.

Using postal vehicles as the basis for a serlsor network taps the principles of a concept
that has been called 'telecartography' or 'geoinformatics', terms meaning the tying
together of measurement data witll their origin location. While the technical literature
suggests that sensor data gathered ill a random, unplanned Inanner and labeled with
location and tinle coordinates can produce some useful nlapping over time, postalbased
fixed-route data collection would produce more useful data more quiclcly and
with greater efficiency. Furthermore, on fixed routes, data from the same locations are
collected daily for a large number of locations, reducing potential sources of error and
increasing opportunities for productive data analysis.

Hardware for such a system would consist of a detector to measure some signal or
paranleter, along with a transducer to convert the results into an electrical signal tllat
can be digitized. The system would require a means of getting the nleasurable quantity
to the detector, for exanlple an air funnel for atmospheric sensing, or a suitable radio
antenna for radiofrequency detection. In addition, a system to transmit the data, a
source of power (either batteries or access to vehicle power), a moullting interface, and
packaging would be needed. Electrical power requirements are nl0dest and would not
interfere with vehicle operation. The equipment would be required to operate over a wide
range of operating conditions, and should be designed so as not to require any attention
fronl the letter carrier/driver. Location- and tillle-stamping of the datastream would be
Sensor network services using the postal delivery vehicle fleet 367
368 Reinventing the postal sector in an electronic age
accomplished with a global positioning system (GPS) device; such devices are now inexpensive
and highly portable. Data transmission can be performed with a wireless device
analogous to a 'smart' phone.

Commercial sensors exist for the applications highlighted in this chapter.! Off-the-shelf
multispecies gas analyzers exist that can detect a wide range of atmospheric constituents
and substances at tiny concentrations.2 Chemical and biological sensors are increasingly
commercially available to detect a wide variety of compounds.

Access to GPS data makes it possible to select a variety of criteria by which data are
collected, and how data are uploaded. For some applications, such as for homeland security
or detecting chemical releases, the primary value of the system will be in providing
an alert mechanism. For others, such as weather data, continuous data collection may be
best. Still other applications may benefit most from an evenly spaced geographic distribution
of data points. For still others, measurements or data transmission is desired only if
some specified contingency occurs. The following list identifies several ways in which data
can be collected and transmitted.
⬢ continuous/real-time data exchange;
⬢ continuous data exchange with lagging buffer;
⬢ periodic measurement (time sampling);
⬢ periodic or opportunistic transmission (burst);
⬢ geographic sampling (location sampling);
⬢ sampling with geographic constraints;
⬢ sampling with time constraints;
⬢ end-of-day or end-of-route upload;
⬢ alert or alarm;
⬢ contingency measurement (triggered by an external event or another measurement);
⬢ contingency transmission (triggered by an external event or a measurement); and
⬢ manually/remotely triggered measurements.