In this paper, we consider fault detection, isolation and reconstruction problem for descriptor systems with actuator faults and sensor faults, respectively. When actuator faults exist in the system, the fault detection and isolation (FDI) problem is solved through an unknown input observer regarding remaining faults excluded a specified fault as unknown inputs. Whereas, in existing sensor faults, the fault detection is only achieved by the unknown input observer and residual signals. Since the derivative signal of sensor fault is generated in the error dynamics between the actual system and the derived observer. The main objective of this work attempts the reconstruction of the faults. The reconstruction can be achieved by sliding mode observer including feedforward injection map and compensation signal. Finally, the isolation problem of sensor faults is solved by reconstructing all of the faults.

No abstract

Generally, tracked vehicles are used in military, agricultural and recreational applications where terrain conditions are improper or unpredictable. Tracked vehicles are better than wheeled vehicles due to the larger contact area of tracks providing better floatation and traction at various ground conditions. In this paper, the path tracking method is proposed for tracked vehicle driving on the deep-sea soft cohesive soil. Firstly, to generate the vehicle's motion following the specified path, the vector pursuit algorithm is used. Next, to drive the vehicle as the generated motion, the control method based on the relation between the traction force and the track slip is proposed. Finally, to prove the performance of vector pursuit algorithm, the numerical simulation is done. Also, to directly induce the relation of traction force-slip, the draw-bar pull test is executed.

A self-propelled miner for deep-seabed manganese nodules was developed for the purpose of evaluations of mining performance and viability for scale-up. The test miner crawls on the seafloor by two tracks and separates the nodules from seafloor by hybrid pick-up device. The weight and size are 10tons in air (5tons in water) and 5m(L)×4m(W)×3m(H), respectively. It is operated electro-hydraulically in real-time via umbilical cable. Software’s for real-time remote control, monitoring and database systems were developed as well. Shallow water tests of the test miner were performed in 100m water depth of an inshore condition site. Glass beads (d = 19mm, m = 10g/ea) were made for artificial nodules instead of real nodules. Non DP barge (W = 19m, L = 51m) was used as surface unit. The sea tests showed that the fundamental performance of the test miner is confirmed and at the same time functional modification and improvement in sensing and measurement system are addressed. This paper describes about the development methods of the test miner and lessons from the sea tests.

No abstract