Project: Dynamic Assistive Information Systems (DAISY)

Reference: STP 14/6/45

Last update: 10/04/2013 15:03:24

Objectives

Enhance independence of people requiring pre-and in-trip information about pedestrian and public transport journeys.

Confirm benefits to people with learning difficulties within the frame of a larger sample and a greater number of contexts than was possible in the previous feasibility project (DIMPLE).

Conclude pre-development researchg of the video map investigated in DIMPLE.

Research information about pedestrian environments and how to incorporate this into the video map.

Develop multi-terrain application (e.g. building sites) where visual cues can help.

"Industrialise" the DIMPLE software to make it easier for users and travel trainers to develop their own journeys using the software, both in-trip and pre-trip.

Description

Software informing people about their location and details of the trip will be industralised for use in the real enviroment. Evaluation of effectiveness will be carried out for users with learning difficulties to confirm the benefits from such software.

Contractor(s)

University College, London
Centre for Transport Studies , Gower Street , London, WC1E 6BT
020 7679 7009

Contract details

Cost to the Department: £212,000.00

Actual start date: 01 March 2003

Actual completion date: 28 August 2006

Publication(s)

Dynamic Assistive Information System (DAISY)
Author: N. Tyler and M Wainstein, UCL
Publication date: 01/03/2007
Source: csau@dft.gsi.gov.uk

Summary of results

  1. Initial attempts - in the early days of the project - at producing software for DAISY did not produce expected results due to the inadequacy of the methods employed to obtain location information. Preliminary methods tried were based on radio beacons, and earlier forms of GPS.

    Radio beacons were found to be very difficult to connect to and did not give accurate enough location information quickly enough. DAISY would be required to be operational on vehicles such as buses. The time required for handshake protocols in Bluetooth, for example, were highly variable and could be as long as 10 seconds. For example, if the user were travelling at a speed of 10m/s (36 km/h), the location detected by their receiver could be 10 metres away from the current true location. We tried two radio-based technologies: Bluetags and iBeans. The Bluetags used Bluetooth as the main connection device and we incurred the problem just described. The iBeans use a single radio frequency and thus do not have such an extended handshake time, but their range was not really good enough. In both cases for DAISY to work, we would require installation of iBeans or Bluetags at suitable locations along a route, thus rendering the system very limited in terms of flexibility of implementation.

    We then looked at GPS in an attempt to solve this problem. GPS was capable of providing extremely accurate location information - even to a few centimetres, but could not do so reliably. Generally, in terms of reliable information, it could be considered to be accurate on average only to several metres in the travelling situation. Car-based GPS systems use sophisticated mapping algorithms to overcome this problem, but these rely heavily on pre-knowledge of the road network and the knowledge that the car could not be in certain locations (e.g. inside a building).

    Experiments showed that for pedestrian use in a built-up area, GPS would be helpful in a general way, but was quite insufficient for use to give navigational directions. The reason for this is that the accuracy for a pedestrian has to be very high - a location error of even 5 metres could mean that the navigation system believes the user is on the 'wrong' side of the road and errors in orientation could mean that they would be considered to be facing in the wrong direction. Issuing an instruction to 'turn left' could result in the user setting off in completely the wrong direction. So for DAISY purposes the technology available at that time was either very costly or unsatisfactory due to poor resolution on location finding and poor satellite signal reception.

    With recent advances in available GPS devices providing improved signal reception in portable form, a new successful attempt has been made to solve the problem. This involves the use of a strong-signal GPS system, which can operate inside buildings up to a point and inside vehicles, communicating this information via a permanent Bluetooth connection (thus eliminating the handshake protocol problem described above).

    Regarding the hardware on which DAISY would be operated, several options were considered and a number of different devices were explored in order to determine the most suitable for development of the system.

    Until recently, portable devices have been expensive, had low processing capability and were limited by the resolution of their displays. DIMLPE finished by using a small laptop computer, but this would clearly be unlikely to be practical in an implemented for of the system. DAISY started from an even smaller laptop - a tablet PC - which offered touchscreen interface with the user, was lighter in weight than the laptop and was able to communicate with the GPS system.

    However, this was still too large for practical use. The increase in use of mobile phones and PDAs, coupled with the increase in accuracy of location-based technology, such as GPS provided relatively high-powered devices capable of running a wide range of applications. New screens were beginning to become available, replacing the monochrome displays with high-resolution colour displays capable of showing high quality photographs and videos.

    We then explored the possibility of suitably enabled mobile phones. The GPS available on mobile phones was quite rare at this stage in the project and very heavy on battery and memory use. After extensive trials of intermediate versions leading to gradual refinement, we obtained a Nokia 6000 mobile phone. This had Bluetooth capability for connecting to the GPS system and a reasonable screen size. It was also able to display the GPS with a downloaded mapping system and it was possible to enter routes for which the GPS would provide navigation instructions. We therefore set out to write the software to enable the Nokia phone to be used with the GPS and to interface with the methods developed during DIMPLE with the users to show the user appropriate photographs at the correct point along the journey.

    The Nokia phone also enabled us to make use of its in-built camera to obtain the relevant photographs. As a result we developed the software to use the GPS system for a given route. At a certain point the navigation system would switch from the GPS to the relevant sequence of photographs. These could locate the user within about 1 metre and ensure that they are facing in a known direction. Thus issuing an instruction to 'turn left' would be appropriate. The system would then revert to the 'standard' GPS navigation mode. We are also able to take advantage of other features - for example, both the GPS information and the instructions linked to the photographs can be given audibly using the phone's in-built voice system, thus removing the need for the user to be able to read the instructions or to have the phone on display while in GPS mode.

    DAISY has reached its final stage of development as a mobile phone-based navigation system, able to give the users appropriate directions based on predefined routes. Navigation is not just assisted by visual references to actual route but also provides users with pre- trip information and in-trip reassurance.

    - Pre-trip information allows users to preview the entire route before they go out, in order to familiarise themselves with where they should be going. They can run the route on the device over and over again. They can step through the route forward and backwards. This is provided by running the software on a PC. Routes can be set and practiced on the PC before downloading them to the phone for use during the journey itself.
    - In-trip reassurance allows users to know exactly where they are on the route. The system has been designed to also provide information about whether they are in the desired location or not, and what to do next in eve