At https://www.blogger.com/comment.g?blogID=19245896&postID=3474881908092116258 S and Josh made some interesting comments in early April 2016.
S pointed out that the Japanese Ministry of Defense is conducting research on two kinds of Unmanned Underwater Vehicle (UUV).
1. A Large Displacement UUV (LDUUV) "for continuous warning and surveillance of threat and for long range delivery of underwater equipment" [Pete located Urashima Platform or LDUUV being developed by the Japan Agency for Marine-Earth Science and Technology (JAMSTEC). The Urashima Platform is 10 meters long and weighs just over 7,250 kg. There may be a dual-use version used by the Japanese Navy (JMSDF)] and
That said, there are problems that would likely challenge such a system.
1. Probably the biggest problem, target classification in such heavily congested waters would be problematic. Nuclear plant noise would be fairly unique to an opponent SSN, but other noise sources would not be SSN specific or even sub specific, particular if the sensor involved was not acoustic or not cross referenced with other sensors (acoustic, radar, visual) that could help reject friendly or neutral shipping, submerged or not (assuming someone didn't do something snarky like intentionally simulate emissions of neutral shipping).
2. The USN is known to have decoys [some older ones] and torpedo target simulators. They come in various sizes from external counter measures tubes (6"), trash disposal unit (9"), and full up heavy weight torpedo size. The larger models in addition to simulating passive sub noises and responding to active pings with an appropriate counter ping also generate a magnetic field to activate MAD sensors.
3. The [seas near China] area involved is a double edged sword to the PLAN [Chinese Navy]: While it is their back yard and readily accessible to their units and even land based weapons, they are late comers to the party: the USN staked out that turf long ago. It is is likely that many operations laying sensors or cables are under some form of observation at this point, either by MPAs, SSNs, or the likely emplacement of the USN's own sensors on the sea floor. The USN first deployed such sensors half a century ago; it stands to reason that some kind of detection system has been in place. It may be the PLANs back yard, but to a certain extent the USN has a home field advantage. In addition to the USN's long time presence in the area, the US can train and test systems in sonar calibration ranges off its west coast and in the Caribbean out of range of any opponent navy; the PLAN is bottled up and has to test, train, and deploy with the constant threat of observation.
4. While a land based rocket propelled torpedo weapon may by physically possible [is with the ASROC system used by Western navies and Russia and China], the targeting challenges associated with using it based solely on seabed emplaced sensor data make it unlikely to be extremely useful without some additional targeting information from some other platform. The tracks of such sensor are likely not weapon delivery grade to begin with, on top of all the complications mentioned above.
This isn't to discount the utility or wide spread use of a detection system or missile delivered ASW torpedo or nuclear depth bomb (particularly if an area was off limits to friendly subs and made a 'free fire' zone for all submerged targets). This is just to note some of the complications such a system would have to face (and its presumed USN counterpart).