AUTOTRACKER: Autonomous Inspection – Capabilities and Lessons Learned in Offshore Operations Jonathan Evans, Pedro Patron, Benjamin Privat, Nicholas Johnson [*], and Chris Capus [**] [*] SeeByte Ltd, Orchard Brae House, Queensferry Road, Edinburgh, SCOTLAND, UK [**] Ocean Systems Laboratory, Heriot-Watt University, Edinburgh, SCOTLAND, UK Abstract-This paper presents AUTOTRACKER, an autonomous pipeline inspection system that operates as a dynamic mission payload for an Autonomous Underwater Vehicle (AUV). The paper describes the mode of operation, together with the validation & trial operations AUTOTRACKER has undertaken over the years, and how this valuable experience has been fed back into the future development of the system. I. INTRODUCTION Probably the most mature of AUV applications is the use of vehicles for seabed area survey. The main aim of using AUVs is to improve the quality of the survey data (for instance, by decoupling the motion of the sensor platform vehicle from the surface), and to reduce the reliance on costly surface ship support. However, for fixed asset inspection such as export pipeline inspection where the distances involve many tens (or even hundreds) of kilometers, accumulated navigation errors and the accuracy of baseline route information mean that AUVs cannot autonomously swim a pre-programmed waypoint mission. AUTOTRACKER is a vehicle agnostic pipeline inspection payload, which can perform onboard dynamic mission replanning to maintain the AUV and its inspection sensors in an optimum attitude for extended durations. AUTOTRACKER’s innovative system architecture provides: Fully autonomous pipeline tracking. Vehicle platform independence. High-resolution, low altitude (2-6 m), high-speed (2-5 knots). Active tracking of pipeline using various sensors, including: Sidescan; Multi-beam; assisted by Legacy route data Full onboard mission re-planning. The ultimate operational aims of using such as system as part of the offshore inspection strategy are improved field reliability and reduced cost of production. In general, the use of autonomous technology is an enabler to different ways of working. For example: Parallel operations on a ROV (Remotely Operated Vehicle) support vessel: Integration into vessel ROV-based inspection regime AUV runs ahead of survey vessel (e.g. 1 day) 1 st pass of data processed overnight ROV and crew follow-on doing spot-checks on anomaly reports the next day Alternatively, AUVs can often use smaller, non-dedicated Vessels of Opportunity: Very useful following events such as post-hurricane cleanup Also, more routine inspection from smaller, flexible vessels – reducing costs and encouraging more routine planned approach to inspection. The AUTOTRACKER inspection system has been developed and thoroughly de-risked before entering commercial operations offshore by a series of trials on a variety of vehicles. This paper outlines the operation of the system and provides an insight into the development and de- risked process. II. AUTOTRACKER: SYSTEM DESCRIPTION A. Inspection Payload The AUTOTRACKER system is comprised to two groups of software components. The group that executes on the host- vehicle platform is referred to as the AUTOTRACKER inspection payload. The onboard payload components are responsible for the live detection and tracking of pipeline in the data obtained from the various onboard sensors, together with the dynamic (not pre-programmed) mission control of the AUV to maintain the optimum vehicle attitude to the actual pipeline position; regardless of errors in the legacy pipeline route file or the accumulated position error on the AUV’s navigational systems. Most AUTOTRACKER configurations for pipeline inspection use sidescan sonar as the primary tracking sensor as this provides the widest area coverage and is present on most AUVs. On vehicles which support additional sensors such (downward-looking) profiling sonars these can be used for tracking at pipeline-vehicle offsets of near zero which cannot be done with sidescan alone because of the transducer and