Manufacturers and operators of aircraft and equipment produce large numbers of technical manuals for their air and ground crews. These manuals generally cover all phases of aircraft operation - from assembly and disassembly, operation and maintenance of the various systems and subsystems, to storage of the equipment. These manuals are constantly revised as and when the customer makes new procedures, or modifications to the existing models.
Most aircraft manuals are written in Simplified Technical English (STE), which is a controlled language developed to help users understand what they read, avoiding misinterpretations and mistakes. This helps the end users, who have a limited knowledge in English in their task of clarifying the various procedures to follow when operating different aircraft systems. Complex technical instructions may be misunderstood which may lead to accidents. Manuals written in STE will be easy to understand and help remove linguistic barriers. However, the chances of finding human errors is higher as the technical authors have to spend more time in proofreading while creating these technical manuals.
The users were calling for an automated solution for verifying the contents of the manuals and highlighting the sections violating the writing rules and dictionary supporting Simplified Technical English. umlaut engineers and experts from the Technical Publication domain based in India and Germany joined their efforts and together, developed an inhouse software employing AI, ML and NLP to increase the efficiency and accuracy of the process of writing the technical manuals. The tool, named Technical Writing Efficiency Enhancement Tool (TWEET), helps the users to upload the authored documents and check for most of the 53 STE rules, generic English grammar, spelling, and other necessary standards applicable when creating the manuals. The tool is also designed to co-relate callouts between corresponding illustrations & task texts and identify missing callouts. Also, its advanced Torque Extraction feature helps the authors visualise specific task-related torque values from Engineering drawings and Bills of Materials. The solution has helped our customer, a leading Aircraft manufacturer, to improve the authoring process quality and reduce human errors.