1. Repair Requirements Analysis Review
The Repair Team
Jason Chalecki
Samuel Chong
Rachel Goldstein
Mark Marrin
Pat Larkin
Stephen Verleye

2. What am I doing here? What are we doing now? (Current System)
What do we want to do? (Proposed System)
How is this going to work, exactly? (Use Cases)
What do we need? (Requirements)
How are we doing this? (Object Models & State Diagrams)
Are we missing something? (Open issues)

3. Current System Paper-based Manual assignment not dynamic expensive
inflexible
easy interface
Manual assignment
Manual matching of maintainers to tasks
flexible
Not dynamic because permanent annotations can’t be conveniently made at time of repair.
Paper is pricey, especially with editing.
Inflexible because technical manuals can’t be automatically searched.
Paper is familiar and substantial
We would like to emulate the flexibility (or, more appropriately, allow the flexible scheduling options) of the current system

4. Proposed System Speech-based IETM and Workorder navigation Benefits
Portable computer with speech recognition and synthesis technologies
Benefits
editing at time of repair or inspection
search capability
“hands-off” interface
automatic submission of workorders
not flammable
We’re going to offer a system that allows interactive navigation of workorders and IETMs, as well as editing and status updates of workorders.
The IETMs will ideally also be able to be searched (requires capabilities from workflow… offers an interface to searching mechanism)
The really novel part is the “hands off” user mode, allowing mechanics to “hear” the instructions rather than interrupting their work in order to read their task.
Finally, the workorder status can be automatically updated.

5. Use Case Diagram
Several Use Cases of our system, focusing on interactions between the data repository and the mechanic.

6. Use Case: Annotate Workorder
Actors: Mechanic, Lotus Notes DB
Entry Condition: Mechanic is working on an active workorder that needs to be edited
Flow of Events:
Mechanic identifies area to be changed (either sticky status or notes section)
Mechanic makes changes in that area on his/her local copy on PEDD
Mechanic submits changes to Lotus Notes DB
Exit Condition: Modified workorder is in Lotus Notes DB
Exception: DB is down, connection is lost Special Requirements: none
Missing a part…changes sticky status...

7. A Common Interaction: RetrieveIETM
This sequence diagram shows a common speech-based interaction with the IETM database.

8. Use Case: Retrieve IETM
Actors: Mechanic, Lotus Notes database
Entry Conditions: Mechanic is logged in
Flow of Events:
Mechanic issues a speech command to retrieve a specific IETM from the database.
The speech system interprets that command and passes the result to the PEDD.
The PEDD executes a query for the IETM.
The PEDD receives the IETM from the database.
The PEDD insures that the grammar is set to IETM navigation, and displays the IETM.
Exit Conditions: IETM is displayed. Exceptions:
Mechanic doesn’t have access to IETM.
IETM doesn’t exist.
Special Requirements: None

9. Requirements Speech-based IETM and Workorder navigation
Speech Recognition Requirements
User training
90% Recognition rate
3-second maximum latency
IETM and Workorder retrieval and notification abilities
Network connectivity
Appropriate hardware
Java 1.2 with JSAPI extensions
Will be able to speak the textual information in the workorders and IETMs, as well as provide speech-based navigation of these documents.
Speech recognition will require training on each user’s voice, and in return will assure a 90% recognition rate with a maximum 3-second latency between the time the command is issued until execution.
Requirements in the hardware dept. aren’t much, easily covered by other computationally intensive subsystems.

10. Object Model (Simplified View)
A pretty simplified view of our object model, showing the structure of our relevant objects. The speech system is omitted from this model, but is closely related to the PEDD (Portable Electronic Display Device, or “the wearable”). Full details are available in our RAD.
The Notes Database contains the user data, as well as the workorders and IETMs. The workorders are mainly comprised of Stickies and annotations, and have pointers to relevant IETMs. The PEDD has access to the workorders and IETMs via the Notes DB.

11. Open Issues Notification interface not well-defined
Current free-form speech recognition not good
IETMs are really big (no, REALLY big)
Authentication standards have not yet been established

12. Got Questions?