JMab
Augur
Maybe you've heard of it, but there was a Kickstarter for a project called CLANG by the sci-fi writer Neal Stephenson and some of his buddies. The idea was to create a realistic medieval fencing simulator and then wrap an RPG around it later on. They were thinking along similar lines, that keyboard and mouse wouldn't be sufficient, so the plan was to create a custom controller that looked similar to a lightsaber handle. Player could hold it like a sword hilt and perform realistic moves which the computer would capture and translate into sword movements. They raised like half a million dollars, but failed to convince anyone else to sponsor it, so as far as I know it's canned or at least on the shelf now.
The best current hope for interesting melee combat is a game called Kingdom Come: Deliverance from a Czech studio. Maybe you heard of it as well. The devs claim the combat is modeled on real historical fencing techniques, and they brought in fencing experts to help them create the system. The game is currently being developed, so they haven't revealed completely how it's going to work yet. Based on what was revealed so far, it will have inverse kinematics (swords will stop when hitting other objects), 3 basic sword attacks (slash, thrust, and pommel strike) that can target 6 different body areas, so 18 overall, and a parrying/prompting system. The latter will require the player to actively parry an enemy strike in time, which will initiate a short period of "bullet-time" during which there will be prompts on the screen for various follow up moves. At this point, it's hard to tell how well it will work exactly without seeing it in action, but the guys behind this game have made some pretty good realistic combat systems in the past (Operation Flashpoint, Mafia).
Yeah, I discovered CLANG about a month ago, when it had gone quiet, prior to it being cancelled. My thought at the time was similar to one of Neal Stephenson's points in his cancel announcement - they seemed to prioritise historical accuracy too much over pure and simple fun of swordfighting. I think the focus for games needs to be fun gameplay number one, with the mechanics built to support the fun.
I'm following KC: D a bit too, I think those guys will get it right. If the game is successful, I think it will raise the bar for melee-based games. It should make it less acceptable for other devs to just have a melee attack button anyway. Different attack types (slash, thrust, etc), various hit targets and targeted parrying/blocking should be a minimum. Not sure about the bullet-time thing though - hope it doesn't just trigger a quick time event sequence. With mostly, or better yet fully, dynamic animation, I would have thought you should be able to trade blows in realtime.
I hope that with KC: D that melee combat still requires thought and timing when you reach higher levels. As opposed to Skyrim, where you quickly become a tank that just needs to spam the attack button to roll over all foes. They seem like pretty clued up guys, so I have faith. I loved Operation Flashpoint in it's day, I thought it was ahead of it's time. I remember them patching in hardware transformation and lighting, which was a big thing to do at the time. Made me think that they like pushing the technical edge, so I hope that KC: D carries the banner forward from Mount and Blade to continue the incline of melee-based combat!