Home    Products    Services   About Us    Contact Us  
ISE Fully Submersible ARCS

AUV Fiber Optic Cable Laying

From Concept To Reality

Phil Hartley & Bruce Butler
ISE Research, Ltd.


The development of autonomous underwater vehicles (AUVs) has been progressing in recent years to a point were some specific applications of the technology can now be applied. It has been widely accepted that AUVs will perform an important role in the future in areas such as deep ocean surveying and military applications. This paper will address an emerging application, fiber optic cable laying. ISE Research Ltd. has completed a series of contracts for Defense Research Establishment Pacific, a Canadian Department of National Defense research institute, which demonstrated the ability of an AUV to autonomously lay fiber optic cable. At the completion of the sea trials, over 30 kilometers of cable had been laid, the longest single cable being 12 kilometers. This paper will address the problem of AUV cable laying from conceptual design through to cable laying sea trials. Follow-on work to develop an AUV capable of performing long range cable laying will be described.


As commercial, oceanographic and military work continues underwater, there is a need for high speed, high quality and secure data links between remote sensor sites and data recording or processing locations. Fiber optic cables are one of the best suited transmission mediums for this application, however installation costs can be very high. These high costs can be compounded when installation is required in areas with extreme water depths or where environmental conditions make normal laying techniques difficult.

One approach at reducing the costs associated with a cable link is to use a very small, low cost fiber optic cable. While the cost of the cable itself can be reduced by using a small, simple design, other installation problems can be introduced. Some of these problems include:

a) cable strain caused by vessel motion during laying;
b) excessive cable in the water column in deep water locations; and
c) the inability to lay the cable over the bottom when operating in areas with unknown bottom topography.

One technique to minimize these problems is to lay the cable from an underwater vehicle. A cost effective approach would be to use an AUV to lay cable close to the bottom regardless of the surface conditions and with minimal support crew.

Defense Research Establishment Pacific in Esquimalt, British Columbia, contracted ISE Research to conduct a feasibility study to determine if the ARCS AUV, owned by the Canadian government, could be used to lay fiber optic cable. The results of the study were positive. Additional contracts were then awarded to allow ISE Research to begin the development of a complete cable laying system.

The cable laying system development contract took place over a 1 year period from January to December 1990. Work carried out included:

a) the design and fabrication of a 50" long hull payload section to house the dispensing system;
b) cable deployment and buoyancy compensation equipment;
c) selection and acquisition of the fiber optic cable; and
d) the subsequent sea trials and laying of approximately 30km of cable.

The cable laying contract focused on the requirements associated with fiber optic cable laying, relying on the previously established abilities of the ARCS to carry out the normal autonomous vehicle tasks such as control and navigation.


System Design Considerations

With the assistance of it's teaming partner Rockwell International, ISE Research compiled a requirements list necessary for mission success. Most of the requirements are directly or indirectly related to the physical parameters of the fiber optic cable.

Cable size, weight and strength are the most important consideration. Cable diameter determines the amount of cable the vehicle can carry, placing an upper limit on mission range. It is desirable to provide the longest range possible to provide cost effective laying operations, thus a very small diameter cable was required. Unfortunately, small cables tend to have less strength. The desire for the smallest size of cable was carefully weighed against reduced strength. The cable must be strong enough to support it's own weight when suspended in the water column and be able to withstand normal handling, but must also have a specific gravity low enough not to cause vehicle buoyancy problems. The cable must also be heavy enough to allow it to sink to the bottom quickly, to avoid excessive lengths of cable from being suspended in the water column.

A design goal for the development system was a cable length of 15 kilometers. A buoyancy compensation system was required to ensure that the ARCS would not become too buoyant as the cable was dispensed. An excessively buoyant vehicle would result in an increase in drag as the hydroplanes attempt to maintain vehicle pitch and depth.

The cable must also be resistant to "hockling", or the formation of loops after being dispensed. If a hockle is present and the cable is tensioned, the glass fiber will likely be broken. For a cable to be hockle resistant, it must have very low residual torque. This places a very stringent requirement on both it's design and manufacture.

Cable pay-out tensions must also be carefully defined. The tension must be high enough to prevent the cable from self dispensing when suspended over the bottom, but not high enough that the vehicle's maneuverability is impaired.

As shown below, the ARCS is configured around a single 27" diameter cylindrical pressure hull. A free flooding section is located aft for housing the propulsion motor and control surfaces. A 50" long free-flooding fiberglass hull section was built to provide space for the cable dispensing and buoyancy compensation equipment. The payload section was inserted between the aft pressure hull bulkhead and the free-flooding motor compartment. A discharge tube was designed and installed to direct the fiber optic cable from the payload area to aft of the propeller.


Cable Selection

A survey was conducted to identify candidate fiber optic cable designs suitable for this application. The cable requirements included:

a) low loss, < 0.25 dB/km;
b) good abrasion resistance;
c) high break strength;
d) low specific gravity, down to approximately 1.5;
e) low torque, < 0.20 in/lb; and
f) low rotation, < 0.25 deg/ft

Of the cables studied, the two most promising candidates were selected and various lengths of each purchased. The cables chosen were: E-glass with a hytrel jacket, and a steel-tubed fiber with FRP and a hytrel jacket. Specifications of the cables purchased are as follows:

Cable Specification E-glass Steel-tubed
break strength 200 lbs. 250 lbs.
working strength 70 lbs. 75 lbs.
specific gravity 1.54 2.2
weight in water 6.2 lbs./km 8.7 lbs./km
diameter 0.067" 0.075"

The cable lengths were wound into cylindrical spools 21" diameter and 12" long, using an orthonormal wind with an internal peel. This winding arrangement provided a very high volumetric efficiency (90%) and would allow the joining of multiple cable spools in future vehicles when longer cable lengths are required.

Appropriate glues were used to hold adjacent layers of cable together during the winding process. The type of glue selected would also determine the resultant pay-out tension. The desired pay-out tension was based on the water depths anticipated combined with the specific gravity of the cable. A design goal of less than1 pound tension was established.


Buoyancy Compensation

As previously mentioned, it is desirable to have a lightweight cable when considering vehicle ballasting. For the most efficient operation of the vehicle, the cable should be neutrally buoyant, so trim and ballast would not change as the cable was dispensed. In practical terms, the limited payload volume available on the ARCS was not sufficient to house a neutrally buoyant cable. In addition, such a cable would not sink after being discharged. A compromise was made between the cable's sink rate and the requirement for some form of buoyancy compensation system.

An evaluation of the longest cable length desired for the present ARCS vehicle (15 km) and the wet weight of the cable (6.3 lbs./km) revealed that without a compensation system, the ARCS would be 94.5 pounds positively buoyant after dispensing the cable. While it would be conceivable for an AUV to operate under this condition, the resulting vehicle drag would increase.

A hard ballast system was designed which allowed water to flow into pressure resistant tanks as the cable was dispensed. The rate of water in-flow was regulated by the increase in buoyancy of the vehicle. The design goal was to keep the ARCS within plus or minus 2 pounds of it's initial buoyancy.

A number of active control schemes were evaluated to control the water in-flow. These included monitoring the angle of attack of the control surfaces, measuring the length of cable dispensed, and comparing the actual vehicle depth with the desired depth. All of these concepts were dependent on the vehicle's control system for operation. It was decided that a passive system would be simpler and more effective.

The design implemented passively monitors the dispensing system's buoyancy and controls the water flow into the tanks with a solenoid valve as the assembly becomes positively buoyant. Using this approach, the dispensing system is self adjusting and that no ballast adjustments are required when cable packs of different length or cable type are installed. The design is also such that as the cable is dispensed and water flows into the tanks, there is minimal change in the difference between the dispensing system's center of gravity and center of buoyancy, resulting in no effect on vehicle trim.


Cable Discharge System

A study was conducted to evaluate possible cable payout systems. These systems would ensure that the cable would be dispensed free of hockles, loops or kinks which could cause failure of the optical fiber. In addition to dispensing the fiber intact, the system must also dispense the correct amount of fiber based on the distance traveled over the bottom.

Both active and passive systems were investigated. Active systems were not chosen due to complexity, system inefficiency due to power consumption, and the possibility of damage to the glass fiber if a double loop or hockle occurred during discharge. A passive design was adopted, relying on the vehicle's forward motion to pull the fiber from the cable pack. The type of glue used to bind the individual wraps of cable together within the pack sets the tension required to pull the cable from the vehicle. This technique ensures that the correct amount of cable will be automatically dispensed from the vehicle provided that the vehicle is operating relatively close to the ocean bottom.

A series of mechanical components were designed to guide and dispense the fiber after it is pulled from the cable pack. The design of the components was such that hockles, if present, will be handled in a manner that will not cause damage to the glass fiber.


Preliminary Testing

Peel-out tests were performed to ensure that the dispenser assembly would properly dispense cable, and to quantify the peel-out tension.

To test the dispensing system, the cable pack, dispenser assembly and payout tube were assembled, and dry and wet peel-out tests were performed at a simulated vehicle speed of 4 knots. The angle of the cable leaving the payout tube was also varied, to simulate different vehicle altitudes.

The peel-out tension of the steel-tubed cable was measured to be 0.75 pounds at a simulated vehicle speed of 4 knots. Based on this figure, the vehicle could operate at an altitude of nearly 400 feet before self-dispensing would occur. No hockles or other problems were encountered.

The E-glass cable had a measured peel-out tension of 0.63 pounds, corresponding to a maximum operating altitude of 800 feet. This cable was found to be quite susceptible to kinking during both the winding and dispensing processes.


Cable Laying Sea Trials

During the fall of 1990, sea trials were undertaken to:

a) determine the effect of cable dispensing on vehicle dynamics;
b) evaluate the cable dispensing system during different maneuvers;
c) evaluate the buoyancy compensation system during cable dispensing; and
d) quantify the characteristics of dispensed cable.

Dynamics of the ARCS vehicle have been measured on several previous occasions. During cable laying trials, vehicle attitude and control surface positions were logged by an onboard data recorder while the vehicle executed a series of two and three dimensional maneuvers. No effects on vehicle dynamics were detected, primarily due to the low cable tension.

To evaluate the cable dispensing system, video cameras and recorders were installed in the vehicle to record cable dispensing. One camera was mounted to view the interior of the cable pack, and another mounted externally to observe the cable as it exited the vehicle. Post-mission analysis of the video provided a means of determining if and where hockling and/or kinking occurred, as well as the angle of the dispensed cable relative to the vehicle during maneuvers.

To determine the effects of the dispensing process on the cables used, an Optical Time Domain Reflectometer was used. At the beginning of each dive, the ARCS was lowered into the water near shore, the cable end pulled from the ARCS and delivered to the OTDR operator on shore. The operator prepared the cable end and measured both attenuation and total length. During the dive, these properties were monitored in near real-time. At the end of the mission, the cable was cut at the vehicle, and final OTDR measurements were made. Neither type of cable showed any significant change in attenuation due to the dispensing process.

A total of 22 cable laying missions, with both types of fiber optic cables, were performed over a 3-month period. Cable lengths laid varied from 200 meters to 12 kilometers. The E-glass cable performed poorly overall, as it was fragile and kinked during both handling and dispensing. The steel-tubed cable performed very well. A kinking problem with the steel-tubed cable early in the program was traced to a solvent used in the glue. The solvent, methyl-ethyl ketone (MEK) temporarily broke down the epoxy matrix in the cable. Switching to a different glue resolved the problem.


Future Directions

There currently exists a need for a larger, long range AUV capable of laying fiber optic cable over hundreds of kilometers. Such missions pose unique problems in hydrodynamics, navigation, obstacle detection and avoidance, control, fault tolerance, and energy storage. ISE Research is presently working on a number of contracts for DREP which are directed at the development of a long range cable laying AUV. The goal is to design and build a system with a range of 300km while laying cables up to 130km. The completion of this system is scheduled for early 1994.

Constrained and free-swimming trials are being performed with the ARCS to generate enough data to accurately model vehicle dynamics, the goal being the design of a larger vehicle hull. A preliminary vehicle design is underway.

The problems associated with long-term autonomous operations are being addressed. A navigation system which will operate anywhere in the world with a range of 300km is being developed, using a Honeywell Inertial Navigation Unit and a Doppler sonar as primary sensors. Present day obstacle detection/avoidance systems are not sufficiently developed for autonomous operation envisioned for this program. ISE Research is being funded to aid in the development of a system suitable for use in an AUV.

A new AUV computer control system is currently under development. Software design uses Object Oriented Programming (OOP) techniques, and a real-time pre-emptive scheduler developed by ISE Controls Group. The hardware platform will be based on the GESPAC G-96 bus, and will use the Motorola MC68030 as the main processor. Fault tolerance will be designed into the control system to ensure successful missions. This system is scheduled for completion by the end of 1991.

The existing ARCS vehicle uses either one or two banks of rechargeable NiCad batteries, each of which allows operations for 5 hours at 4 knots. Energy systems with significantly greater capacity which are being investigated include silver cathode batteries and fuel cells.

A phased approach has been taken with the present development contracts, intending to demonstrate the ARCS vehicle with it's new control and navigation systems in early 1992-3. Following the testing of these subsystems, they will be transferred to the new vehicle for completion and testing in early 1994.

In addition to the intended cable laying use, the technology developed during this project has other interesting applications. The concept of an underwater vehicle carrying and dispensing a small fiber optic cable is of significant importance when considering real time control of systems in extreme water depths. By utilizing the techniques demonstrated during this project, vehicles can be conceived which will offer real-time control to full ocean depths without the necessity of a large surface winch.


In less than 12 months, ISE Research, with assistance from Rockwell International, designed, fabricated, installed, and field trialed a fiber optic cable laying system for the ARCS AUV. Sea trials demonstrated that an AUV could autonomously lay different types of fiber optic cable without compromising cable integrity. During the final mission, nearly 12 kilometers of cable was laid in less than three hours.

ISE Research is presently working on the design of a cable laying AUV with a range of 300 km. The completion of this system is scheduled for 1994.

Applications of this technology will be of great use in systems intended for extreme water depths.


The authors would like to acknowledge DREP for their continued support of this program. We would also like to thank Stanley Stone of Rockwell International for his assistance and perseverance during sea trials, often in less than ideal weather conditions