The RADS Page
RADS : The Route ADvisory System (1997-2002)
Leong Hon Wai,
School of Computing, NUS.
(See also latest news here.)
RADS -- Route ADvisory System
RADS is short for
Route ADvisory System.
RADS was a research and development project led by
Prof. Leong Hon Wai of the School of Computing (SOC), of the
National University of Singapore (NUS). The development of RADS
and its associated enhancements and related research spanned five
years 1997-2002 and involved many students, many of them undergraduates
of NUS School of Computing.
The RADS system enables Singapore commuters to
plan the best route on all public transportation
(all buses and subway) between
any two points in Singapore.
RADS has many innovative features that distinguishes
it from other route planning systems. For example, RADS
- seamlessly combines all modes of transportation (all public buses and subway),
- allows user to optimize fare or time or some combination,
- takes into account estimated walking time to the intermediate stops,
- accounts for the estimated waiting times at the intermediate stops,
- is dynamic and will give up-to-date real-time optimal routes
if it receives real-time information on bus arrivals.
Versions of RADS
At the moment, RADS comes in several options:
- Windows stand-alone application,
- experimental web-based service with simple text output,
- experimental WAP-based service activated via Phone emulator.
Windows stand-alone version:
The Windows version of RADS come integrated
with a rudimentary GUI-driven GIS/map interface
to view a (skeletal) map of Singapore
with partial mockups with several landmarks
and several major roads for demo purposes.
It also shows the locations of all the bus stops (as of about 2001).
(This research project did not have the funding to support the
purchase of the Singapore digital map data from SLA.)
The RADS GUI/GIS supports some of the basic GIS functions
like zooming, panning.
This GUI of the Windows version of RADS
- allows users to select start/end points (and preferences) and plan their routes,
- displays the planned route,
- allows easy re-planning of the routes using new preferences,
- allows users to check bus route information (including map of the route).
Web and WAP versions:
These versions are experimental trials that deploy the route planning
engine for the web and wap. Both versions do NOT come with digital maps.
A research server in the NUS School of Computing
is used to provide this experimental route planning service for web/wap.
This server receives the start/end points and planning preferences
and will compute the planned route, and
return the planned route in xml/wml formats.
For the web version of RADS, the user will have to hand code the locations
of start and end points and the planning preferences,
to be submitted to the server.
(Note: A web form can be created to simplify this process a little.)
In the WAP version of RADS, we make use of a WAP phone emulator that provide
simple GUI-like features to select start and end points and planning
preferences. The planned route is returned in wml that is
displayed on the emulated phone.
The output comes in three levels of details.
The default is very brief desciption of the route: bus service, MRT stops.
The intermediate level (activated by choosing "more" of the phone)
shows more details of bus-stops. Finally the verboise level gives a lot
of details bus/MRT service to take, the stops/stations to board
and where to alight, and so on.
Scalability Issues:
Extensive performance benchmarking has been performed on this
experimental route planning service in order to ensure
scalability of this web service.
(Note: Because research funding for this project has ended,
this RADS web service is turned off (by default).
But it can be re-activated if there is a special request.)
Note 2: (Added in 2008)
The RADS web service is coming alive again.
In collaboration with IDA, RADS will be provided as a web-service, codenamed the
IDA-NUS Free Trial RADS Service.
IDA will work it her partners to provide a Web 2.0 Google Map interface
for all web-users to select their travel plan (source, destination, preferences, etc)
while the route planning and optimization will be done by
the RADS web-service of our group.
For early preview of things, please view the following
Key Technologies used in RADS
While the problem of finding a bus/MRT route between two bus stops
is a relatively trivial algorithmic problem, it solves only the
static problem (namely, is there a route between these two points).
In fact, this problem can/should be better solved via
pre-computation of the routes and then doing a table look-up.
However, to find an optimal route for
different user preferences
and that integrates walking time, and
waiting time at the bus stop
is a more complex optimization problem and requires considerably more
sophisticated algorithms. In fact, RADS can also take advantage
of real-time arrival schedule information to give up-to-date real-time
optimal routes.
The key technologies developed by the research group
and deployed in RADS is a novel route optimization engine
that we call RADS-MM.
The RADS-MM route optimization engine incorporates an
algorithm that computes shortest travel
plan using public transportation (buses and MRT).
Technically, the RADS-MM route optimization engine solves
a dynamic multi-modal, multi-criteria shortest path problem.
Other RADS-related Development Efforts:
There are other R&D efforts that are related to RADS, but are
either less matured or have not been integrated into the RADS system.
We list some of them here for information (and in case other parties
are interested in them).
- Proximity Search (RADS-PS):
Extension of RADS to incorporate location-based services
(search for "interesting services/places") nearby to and
along the planned route.
This RADS-PS feature allows users to see interesting
places and/or events happening at locations near along the planned route.
Research on the technology for RADS-PS has been done and
RADS-PS is a relatively mature and stable technology.
However, RADS-PS has NOT been integrated with RADS
since we do not have real data on location-based services.
Testing has been done on randomly generated data of
location-based services.
- Route Planning for Cars (RADS-CAR):
Extension of RADS to do path planning for cars on the road.
Several routing algorithms for RADS-CAR has been evaluated and
we have development refinement to a selected algorithm.
The testing of RADS-CAR is based on a very old version of the
road map of Singapore.
RADS-CAR has not been fully integrated into the main RADS system.
- Routing Engine Version 2 (RADS-MM2):
Version 2 of the multi-modal engine will have several new
features -- better user customizations, accurate reflection
of the bus fares (including fare rebates), and
computation of multiple shortest routes in one go.
We have designed an algorithm for some issues, and have
partial implementation of the above feature enhancements.
They were in the testing and evaluation phase.
This has NOT been integrated with RADS.
Acknowledgements and Credits:
The research and development of RADS span many years (1997 to 2002)
and is the combined efforts of many of my students who
have worked on different parts of RADS.
They includes: Foo Hee Meng (past project
manager role), Lao Yizhi (RADS-MM), Lam Nah Peng (RADS-Dyn),
David Ong (project manager for RADS-GUI and RADS-GIS),
Desmond Kao, Esther Choa, Stevenas (RADS-GUI),
Li Qiming, Liang Zhe, Sun Guangyu (RADS-GIS), Michael Teo (RADS-Web),
Ho Ngai Lam, Francis Ng Hoong Kee (RADS-PS),
Yuan Yi, Yong Chin Loung, Hu Xiao (RADS-GUI2),
Gokul Poduval, Sharma Vikas, Ngai Nyuk Lan, Dorothy (RADS-WAP),
Phan Thi Xuan Linh (RADS-CAR),
Tan Meng Mau, Guan Bo, Hieu (data),
Kal Ng Yen Kaow (various assorted enhancements).
| Leong Hon Wai
| Dept of Computer Science
| School of Computing
| NUS |
(Started: Jan 2002, Updated: July 2008)
This document, RADS-Info.html, has been accessed 99 times since 25-Jun-24 11:57:13 +08.
This is the 4th time it has been accessed today.
A total of 56 different hosts have accessed this document in the
last 150 days; your host, 3.145.57.41, has accessed it 1 times.
If you're interested, complete statistics for
this document are also available, including breakdowns by top-level
domain, host name, and date.