Date Range: 02/6-12/2019
Total Hours Spent: 7
Overview
As I alluded to in my previous milestone update, this time of the semester is busy, busy, busy. We've been scrambling to finish assignments and study for midterms but, in the midst of all the chaos, I've also managed to plant a flag at another solid milestone for Kendo Gakko. As usual, I've been keeping track of our GitHub project board to continue creating, updating, merging, and organizing tickets to keep our backlog as unambiguous as possible. This is, and shall be until the final build has been submitted, an ongoing process given the nature of this project and our team's inherent lack of experience building experiences with A-Frame and JavaScript. That being said, the team has fallen into a good rhythm and, despite our heavy work load, has maintained a pace of steady, incremental progress. My last milestone marked the introduction of Oculus display and Touch controller support and now, building on that, I've managed to bridge the gap between the Oculus Rift and physics systems to integrate real time physical interactions between the user and the virtual objects around them.
Challenges & Successes
The challenge of time management is ever-present here in the IMD program as I may have mentioned, perhaps a little too passionately, last week. This sprint was no different but the satisfaction of integrating physical interactions, still with challenges of its own, helped balance out the stresses of time.
I focused on making incremental progress with the end goal being the ability to pick up the proxy-model shinai and smack around some objects with it. I quickly learned that, as far as I know so far, the physics system we're using with A-Frame does not like parenting hit boxes to other hit boxes, which may present some challenges later on in the project. Investigations ongoing. As a result, I had to fiddle with the hierarchy when trying to parent a hit box to the hand-controls models before landing on a configuration that allowed me to secure a temporary proxy sword to the right hand model (see video below). With our brand new red stick I introduced the ability to smack some objects around with something that resembled a sword, success! Unfortunately, in the process, I discovered that our physics system struggles to simulate the physics collisions of a fast-moving body; hitting a block at high speeds sends it hurdling off in the opposite direction of the incoming strike. Luckily, our end-product does not rely on the realistic physics of high speed collisions but instead only needs to register the collision, which we can. So, we're good.
I originally rigged placeholder geometry to the model for the right hand control that allowed us test out collision detection and physical interaction between objects.
The next step was to figure out the math behind using our physical interactions to pick up a modeled shinai proxy and have it behave as you'd expect when holding a sword. Again, this was a game of parenting hierarchies and fiddling with position offsets to make sure that when the model is grabbed, it will end up in your [virtual] hand, not your face or floating somewhere off in space, and naturally follow your movements. Similarly, when the model is released from your grip, the sword should leave your hand with as little 'jank' as possible and fall to the floor - not through it.
Next Steps
We will putting together a brief alpha-build presentation for this Friday to show off the progress we've made so far and reassure our inquisitive investors that we'll be able to deliver on our original designs. Afterwards, we'll be aiming to get a solid chunk of work done over the reading week before coming back for milestone 5 on February 26th. The team will be using the extra time away from lectures to tackle features like user-device detection, to automatically direct uses to the appropriate experience according to their platform, and a bunch of UX design work for mobile interfaces and graphics to be used in VR.
Commentaires