Collaborative virtual environment
More actions
A collaborative virtual environment (CVE) is a virtual world shared by several participants over a computer network, in which each person is represented by an avatar and can communicate with the others and work on the contents of the world together. In a 1998 review, Elizabeth Churchill and Dave Snowdon defined a CVE as "a distributed, virtual reality that is designed to support collaborative activities", providing a graphically realised digital landscape in which multiple users "can interact with each other and with simple or complex data representations".[1]
The term comes from academic research of the 1990s, where it sat between virtual reality and computer-supported cooperative work (CSCW). Research systems such as DIVE at the Swedish Institute of Computer Science, MASSIVE at the University of Nottingham and NPSNET at the U.S. Naval Postgraduate School addressed problems such as how to keep many distant users in step, how much each person needs to see and hear of everyone else, and how avatars should show what their owners are doing.[2]
Definition
Steve Benford, Chris Greenhalgh, Tom Rodden and James Pycock, writing in Communications of the ACM in 2001, described CVEs as technology that "aims to transform today's computer networks into navigable and populated 3D spaces that support collaborative work and social play". In their account, participants are given graphical embodiments, called avatars, that convey "their identity, presence, location, and activities to others", and they communicate through a mix of audio, video, graphical gestures and text. The look of both the world and the avatars can vary widely from one system to another.[2]
Churchill and Snowdon noted that CVEs were being used both by geographically separated collaborators and by people in the same room, and that systems differ in how detailed their data and embodiment representations are and how much interactivity they support. They argued that a system meant to support collaboration should be designed around the tasks to be done and the social and cognitive characteristics of its intended users.[1]
The concept overlaps with several neighboring terms. A CVE does not require a head-mounted display: the DIVE platform, for example, could be used on an ordinary desktop PC, immersively with a head-mounted display, or with projected displays.[2] Later research on social VR addresses many of the same questions; a 2026 scoping review describes social VR as providing "immersive, spatial, and three-dimensional social interactions" and examines how it supports remote collaboration.[3]
History
Networked military simulation
An early shared simulated world was built for military training. SIMNET, whose name came from "simulator network", was a prototype network of tank, armored vehicle and aircraft simulators put together by the Defense Advanced Research Projects Agency (DARPA) to train U.S. Army and Air Force personnel. Its program manager, Air Force Col. Jack Thorpe, first tackled the simulation problem in 1978 and extended his studies to the Army after he was assigned to DARPA in 1981. DARPA built a plywood mock-up tank simulator by January 1985, showed a full simulator with crude interactive graphics in October 1985, and installed the first two preproduction units at Fort Knox in May 1986. By 1989 the test-bed linked a central node at Fort Knox to outlying sites in the United States, with three sites in West Germany tied to a node at Grafenwoehr, all linked over 56,000 bit-per-second long-haul lines. The simulators were modular workstations on a distributed network rather than replicas built around one central computer, and every participant in an exercise fought the same battle, in teams against other human teams rather than computer opponents.[4]
Work at the Naval Postgraduate School carried networked simulation toward the general problem of scale. A 1995 paper by Michael Macedonia, Michael Zyda and colleagues described a network architecture that partitioned a virtual environment by associating spatial, temporal and functionally related classes of entities with IP multicast groups, so that each entity's processing and network resources were limited to its own area of interest. The authors described their NPSNET vehicle simulator as, at the time, the only Distributed Interactive Simulation (DIS) protocol compliant simulator using IP multicast and said it was suitable for use over the Internet.[5]
Research systems of the 1990s
Benford and colleagues described CVEs as the result of research interests converging in the VR and CSCW communities. For VR researchers, CVEs extended single-user systems to multiple participants, which suited training, shared visualization and multiplayer games. For CSCW researchers, they offered a shared space that might support social interaction which audio conferencing, videoconferencing and shared desktop applications did not, including crowded settings where tens or hundreds of people form and dissolve subgroups.[2]
The Distributed Interactive Virtual Environment (DIVE), developed at the Swedish Institute of Computer Science, was one of the first CVEs. Christer Carlsson and Olof Hagsand published a description of the platform in 1993.[6][2] Hagsand later described DIVE as a software platform for multiuser virtual environments that "has served as a toolkit for many distributed VE applications", with an emphasis on networking and human-computer interaction and support for autonomous behavior-driven objects, collision detection, audio and 3D navigation.[7] By 2001 DIVE supported real-time network audio, video, text chat, simple graphical gestures and shared whiteboards; its "universe" was made of many separate worlds linked by portals, object behaviors could be scripted in Tcl, and the system was freely available for research use.[2]
Benford and Lennart Fahlén presented a spatial model of group interaction at the 1993 European Conference on Computer-Supported Cooperative Work. The model used the abstractions of aura, focus, nimbus and adapters to control how aware objects in a virtual space are of one another, and the authors framed it as applicable to any collaborative system with a spatial metric, not only to virtual reality.[8] Greenhalgh and Benford implemented it in MASSIVE, a prototype VR teleconferencing system reported in ACM Transactions on Computer-Human Interaction in 1995. MASSIVE let several users talk through any combination of audio, graphics and text over local and wide area networks, with each user's perception of another depending on their relative positions and orientations. The authors reported that early trials showed "the importance of audio in collaborative VR" and raised issues of field of view, navigation speed, quality of embodiment and scalability.[9] Its successors MASSIVE-2 (1996) and MASSIVE-3 (1999) were used for a range of applications, especially public participation in online art and performance.[2] The Nottingham group's work led to the idea of inhabited television, which combines a CVE with broadcast television: members of the public take part in shows staged in a shared virtual world, and virtual cameras capture the action for conventional viewers.[2]
Studierstube, a collaborative augmented reality system, was first presented at the Collaborative Virtual Environments (CVE'96) workshop in Nottingham in September 1996.[10] The Association for Computing Machinery published proceedings of an international conference series on collaborative virtual environments: the third conference, sponsored by ACM SIGGRAPH, SIGCHI and SIGGROUP, was held in San Francisco in September 2000, and the fourth in Bonn in September 2002.[11][12]
Commercial activity ran in parallel. Benford and colleagues noted proposals to extend the Virtual Reality Modeling Language (VRML97) for multiple users, and companies building Internet CVEs either on VRML extensions or on proprietary technology. They cited Active Worlds, whose inhabitants built cityscapes "in a Lego-like manner" and had filled its worlds with over twenty million objects, and the Blaxxun Community Platform, a modular client/server system for running web communities.[2]
Collaborative augmented reality
CVE research also took in augmented reality, where virtual content is overlaid on the real world rather than replacing it. Mark Billinghurst and Hirokazu Kato wrote in 2002 that immersive and desktop CVEs restore some of the spatial cues of face-to-face conversation but require users to enter a world separate from their physical surroundings, whereas AR lets collaborators see each other and the shared virtual objects at the same time. They named Vienna University of Technology's Studierstube project, which used see-through head-mounted displays to let users view 3D models together, as one of the earliest AR interfaces for face-to-face collaboration, along with the AR2 Hockey air hockey game from the Japanese Mixed Reality Systems Laboratory.[10]
Their own 1998 AR conferencing interface showed a remote collaborator as a life-size live video window attached to a physical card, which was tracked with computer vision by recognizing black squares printed on it. In a 1999 user study comparing this interface with audio and video conferencing, subjects reported a significantly stronger sense of presence for their remote partners in the AR condition. The same group's MagicBook let some users view a scene as AR over a real book while others were immersed inside it, each group seeing the other at a different scale.[10]
Commercial and consumer platforms
Sebastian Friston and colleagues wrote in 2021 that social VR systems, long a subject of academic study, were "now attracting increasingly large audiences" on consumer VR headsets.[13] In a 2022 survey of synchronous AR, VR and mixed reality remote collaboration systems, Alexander Schäfer, Gerd Reis and Didier Stricker reviewed 87 systems, 25 of them commercial or professional. They wrote that the COVID-19 outbreak had emphasized the importance of such systems, noted that the IEEE VR 2020 conference was held fully online, with virtual meeting rooms built on Mozilla Hubs, and found that commercial systems usually allowed about 20 to 50 people to share a virtual space at once.[14]
Several large companies built workplace meeting apps on the same model and have since closed them. Facebook released Horizon Workrooms as a free open beta for the Oculus Quest 2 on August 19, 2021, supporting "up to 16 people in VR together, and up to 50 people total on a call", with spatial audio, a shared whiteboard and tracking of the user's real desk and keyboard.[15] In January 2026 Meta said Workrooms would be discontinued, giving users until February 16, 2026, and that it was stopping sales of its commercial Quest models and managed services.[16] In 2025 Microsoft told customers that it would retire the Microsoft Mesh PC and Quest apps, the Mesh website and the Immersive Space (3D) view in Teams as of December 1, 2025, with immersive events in Teams remaining as the replacement.[17] Mozilla announced in February 2024 that it would shut down its hosted Hubs service, a WebXR platform for "private, virtual 3D worlds in your browser", on May 31, 2024, leaving the open-source code to be maintained by the community.[18]
Technical challenges
Scalability and interest management
Real-time interaction among many participants spread over a wide area network makes scale a core technical challenge for CVEs. Bottlenecks include network traffic from movement updates and audio, server processing, the "last mile" connection to each user (in 2001, often a dial-up modem), and the local computer's ability to render the world quickly enough.[2]
The usual response is interest management: each participant receives only the part of the world they are likely to care about. NPSNET divided the world into fixed-size hexagonal cells, with participants sending updates to their current cell and receiving from cells within their area of interest, an approach suited to battle simulations with predictable movement. SPLINE, a system from Mitsubishi Electric Research Laboratories, joined variable-sized "locales", each with its own coordinate system, in flexible ways. MASSIVE-2 used regions whose boundaries could be more permeable in one medium than another, for example blocking audio while letting visual information through.[2]
Benford and colleagues also described using higher-level spatial semantics to manage scale, through the spatial model of interaction. A participant's focus is a field describing where their attention is directed, and their nimbus describes how they project information into space; one participant's awareness of another is a function of the first person's focus and the second's nimbus. A large nimbus can model shouting and a small, narrow one whispering.[2][9] Each object's aura defines the total region within which it interacts, and the model is applied separately in each medium.[9]
Distribution architectures
Benford and colleagues described three basic ways of distributing a shared world, often combined within one system:[2]
| Architecture | How it works | Examples given in 2001 |
|---|---|---|
| Client/server | Each client talks only to a server, which forwards messages to other clients and can tailor traffic to each client's network and machine | MASSIVE, SPLINE and, to a lesser extent, DIVE used servers to coordinate joining a world; the norm for public Internet CVEs |
| Peer-to-peer unicast | Clients send updates directly to each other; most bandwidth-hungry, but avoids server load and usually lowers delay | MASSIVE-1 communication; video streams in FreeWalk |
| Peer-to-peer multicast | The same update is sent once to many peers using a mechanism such as IP multicast | All traffic in NPSNET; all updates in DIVE and MASSIVE-2; audio in SPLINE |
Because wide-area IP multicast was not widely available, some systems added multicast bridging and proxy servers. SPLINE, for example, combined a peer-to-peer core with client/server access for low-bandwidth dial-up users.[2]
Most CVEs show every participant the same content from different viewpoints. Benford and colleagues pointed out that CSCW experience with "What You See Is What I See" interfaces had led researchers to add private views and different views of shared data, and that some CVEs had begun to offer "subjective" views reflecting each user's role. Examples include an architectural model in which different participants see wiring, plumbing or network overlays, and inhabited television, where virtual cameras could be visible to performers but hidden from other participants.[2]
Embodiment and human factors
Benford and colleagues wrote that studying people in CVEs may require methods beyond the perceptual psychology used to evaluate single-user VR. Ethnographic studies of early MASSIVE-1 trials found that even graphically simple avatars could represent participants effectively at times, but that trust broke down when an avatar turned out to be unoccupied because its owner had stepped away. Further studies found difficulties with humanoid avatars: participants assumed others could see objects in their peripheral vision when the system's field of view was in fact much narrower.[2] Billinghurst and Kato identified a related problem in collaborative AR, where head-mounted displays hide the wearer's eyes and so remove gaze as a communication cue.[10]
Applications in VR and AR
- Military training: SIMNET let crews at distant sites train together in one simulated battle, and U.S. entrants used it to recreate the Grafenwoehr range while preparing for the June 1987 Canadian Army Trophy, a competition among NATO armored units.[4] NPSNET-IV was designed for large-scale military training and simulation exercises with thousands of interacting entities.[2]
- Teleconferencing and remote work: MASSIVE was built for VR teleconferencing,[9] and commercial apps such as Horizon Workrooms and Microsoft Mesh later offered meetings in shared virtual spaces.[15][17]
- Shared visualization and design: Researchers expected teams of scientists or decision-makers to share and discuss visualizations, and participants to manipulate shared 3D models while each kept their own viewpoint.[2] Billinghurst and Kato gave the example of a collaborative urban design application in which users at a real table see virtual building models and move them on a physical street map.[10]
- Entertainment and inhabited television: MASSIVE-2 was used for "Out of this World", an inhabited television game show staged for the isea98:revolution digital arts festival, in which the public controlled two competing avatar teams led by immersed actors while four virtual cameras produced a broadcast shown to a theater audience.[2]
- Education: A 2021 open-source toolkit, Ubiq, was built on Unity to provide connection management, voice and avatars for social VR systems, and its authors reported using it successfully in classroom teaching.[13]
Research
Later reviews connect the 1990s CVE work to current mixed reality collaboration. Barrett Ens, Mark Billinghurst and co-authors reviewed three decades of collaborative mixed reality research in 2019 and set out to reconcile that work with CSCW theory, arguing that MR technology had only recently matured to the point where researchers could focus on the nuances of supporting collaboration rather than on creating the enabling technology.[19] Schäfer, Reis and Stricker organized remote collaboration systems into three components, environment, avatars and interaction, and, to their knowledge, found no commercial AR-only system for real-time synchronous remote collaboration at the time of their survey; most commercial systems relied on VR, often with desktop and mobile clients, which raises problems of asymmetric input between users with different devices.[14]
A scoping review by Niloofar Sayadi, Sadie Co and Diego Gómez-Zará, published in 2026, screened 2,035 articles from 20 years of social VR research and identified 62 that addressed collaboration. It found that social VR can support collaboration at three levels: individual perceptions within groups, team dynamics, and VR-specific ways of augmenting users' spaces. The authors called for research to move beyond replicating physical-world interaction.[3]
Notable systems
| System | Developer | Date | Notes |
|---|---|---|---|
| SIMNET | DARPA | Prototype units from 1986 | Networked tank and aircraft simulators sharing one battle[4] |
| DIVE | Swedish Institute of Computer Science | Described 1993 | Multicast-based CVE toolkit, usable on desktop or immersively[6][2] |
| MASSIVE | University of Nottingham | 1995 | VR teleconferencing with the spatial model of interaction[9] |
| MASSIVE-2 and MASSIVE-3 | University of Nottingham | 1996 and 1999 | Regions, online art and performance, inhabited television[2] |
| NPSNET-IV | U.S. Naval Postgraduate School | 1990s | Large-scale military simulation, IP multicast, hexagonal cells[2][5] |
| SPLINE | Mitsubishi Electric Research Laboratories | 1990s | Java middleware; worlds built from linked locales[2] |
| Studierstube | Vienna University of Technology | First reported 1996 | Face-to-face collaborative AR with see-through HMDs[10] |
| Horizon Workrooms | Meta Platforms | 2021-2026 | Quest meeting app; up to 16 people in VR[15][16] |
| Microsoft Mesh | Microsoft | Retired December 2025 | Immersive Teams meetings and Mesh apps[17] |
| Mozilla Hubs | Mozilla | Hosted service ended 2024 | Browser-based WebXR rooms; code handed to the community[18] |
See also
References
- ↑ 1.0 1.1 E. F. Churchill, D. Snowdon (1998-03). "Collaborative virtual environments: An introductory review of issues and systems". Virtual Reality, vol. 3, no. 1, pp. 3-15. Springer. doi:10.1007/BF01409793. https://doi.org/10.1007/BF01409793. Retrieved 2026-10-04.
- ↑ 2.00 2.01 2.02 2.03 2.04 2.05 2.06 2.07 2.08 2.09 2.10 2.11 2.12 2.13 2.14 2.15 2.16 2.17 2.18 2.19 2.20 2.21 2.22 Steve Benford, Chris Greenhalgh, Tom Rodden, James Pycock (2001-07). "Collaborative Virtual Environments". Communications of the ACM, vol. 44, no. 7, pp. 79-85. Association for Computing Machinery. doi:10.1145/379300.379322. https://cacm.acm.org/research/collaborative-virtual-environments/. Retrieved 2026-10-04.
- ↑ 3.0 3.1 Niloofar Sayadi, Sadie Co, Diego Gómez-Zará (2026-04). ""Feeling that I was Collaborating with Them:" A 20-year Scoping Review of Social Virtual Reality Leveraging Collaboration". Proceedings of the ACM on Human-Computer Interaction, vol. 10, no. 2, article CSCW017. Association for Computing Machinery. doi:10.1145/3788053. https://doi.org/10.1145/3788053. Retrieved 2026-10-04.
- ↑ 4.0 4.1 4.2 John Rhea (1989-08). "Planet Simnet". Air Force Magazine. https://www.airandspaceforces.com/PDF/MagazineArchive/Documents/1989/August%201989/0889Simnet.pdf. Retrieved 2026-10-04.
- ↑ 5.0 5.1 M. R. Macedonia, M. J. Zyda, D. R. Pratt, D. P. Brutzman, P. T. Barham (1995). "Exploiting reality with multicast groups: a network architecture for large-scale virtual environments". Proceedings of the Virtual Reality Annual International Symposium '95, pp. 2-10. IEEE. doi:10.1109/VRAIS.1995.512473. https://doi.org/10.1109/VRAIS.1995.512473. Retrieved 2026-10-04.
- ↑ 6.0 6.1 Christer Carlsson, Olof Hagsand (1993-11). "DIVE - A platform for multi-user virtual environments". Computers & Graphics, vol. 17, no. 6, pp. 663-669. Elsevier. doi:10.1016/0097-8493(93)90115-P. https://doi.org/10.1016/0097-8493(93)90115-P. Retrieved 2026-10-04.
- ↑ O. Hagsand (1996). "Interactive multiuser VEs in the DIVE system". IEEE MultiMedia, vol. 3, no. 1, pp. 30-39. IEEE Computer Society. doi:10.1109/93.486702. https://doi.org/10.1109/93.486702. Retrieved 2026-10-04.
- ↑ Steve Benford, Lennart Fahlén (1993). "A Spatial Model of Interaction in Large Virtual Environments". Proceedings of the Third European Conference on Computer-Supported Cooperative Work (ECSCW '93), pp. 109-124. doi:10.1007/978-94-011-2094-4_8. https://doi.org/10.1007/978-94-011-2094-4_8. Retrieved 2026-10-04.
- ↑ 9.0 9.1 9.2 9.3 9.4 Chris Greenhalgh, Steven Benford (1995-09). "MASSIVE: A Collaborative Virtual Environment for Teleconferencing". ACM Transactions on Computer-Human Interaction, vol. 2, no. 3, pp. 239-261. Association for Computing Machinery. doi:10.1145/210079.210088. https://doi.org/10.1145/210079.210088. Retrieved 2026-10-04.
- ↑ 10.0 10.1 10.2 10.3 10.4 10.5 Mark Billinghurst, Hirokazu Kato (2002-07). "Collaborative Augmented Reality". Communications of the ACM, vol. 45, no. 7, pp. 64-70. Association for Computing Machinery. doi:10.1145/514236.514265. https://doi.org/10.1145/514236.514265. Retrieved 2026-10-04.
- ↑ "Proceedings of the third international conference on Collaborative virtual environments (CVE '00)". ACM Digital Library. Association for Computing Machinery. 2000-09. doi:10.1145/351006. https://doi.org/10.1145/351006. Retrieved 2026-10-04.
- ↑ "Proceedings of the 4th international conference on Collaborative virtual environments (CVE '02)". ACM Digital Library. Association for Computing Machinery. 2002-09-30. doi:10.1145/571878. https://doi.org/10.1145/571878. Retrieved 2026-10-04.
- ↑ 13.0 13.1 Sebastian J. Friston, Ben J. Congdon, David Swapp, Lisa Izzouzi, Klara Brandstätter, Daniel Archer, Otto Olkkonen, Felix J. Thiel, Anthony Steed (2021-12-08). "Ubiq: A System to Build Flexible Social Virtual Reality Experiences". Proceedings of the 27th ACM Symposium on Virtual Reality Software and Technology (VRST '21). Association for Computing Machinery. doi:10.1145/3489849.3489871. https://doi.org/10.1145/3489849.3489871. Retrieved 2026-10-04.
- ↑ 14.0 14.1 Alexander Schäfer, Gerd Reis, Didier Stricker (2022-12-07). "A Survey on Synchronous Augmented, Virtual, and Mixed Reality Remote Collaboration Systems". ACM Computing Surveys, vol. 55, no. 6. Association for Computing Machinery. doi:10.1145/3533376. https://doi.org/10.1145/3533376. Retrieved 2026-10-04.
- ↑ 15.0 15.1 15.2 "Introducing Horizon Workrooms: Remote Collaboration Reimagined". Meta Newsroom. Meta Platforms. 2021-08-19. https://about.fb.com/news/2021/08/introducing-horizon-workrooms-remote-collaboration-reimagined/. Retrieved 2026-10-04.
- ↑ 16.0 16.1 Richard Speed (2026-01-16). "Meta retreats from metaverse after virtual reality check". The Register. https://www.theregister.com/2026/01/16/meta_quest_horizon_workrooms/. Retrieved 2026-10-04.
- ↑ 17.0 17.1 17.2 Mary Jo Foley (2025-07-10). "Microsoft is finally nixing its Mesh mixed reality efforts". Directions on Microsoft. https://www.directionsonmicrosoft.com/microsoft-is-finally-nixing-its-mesh-mixed-reality-efforts/. Retrieved 2026-10-04.
- ↑ 18.0 18.1 Ian Hamilton (2024-02-16). "Mozilla Hubs Is Shutting Down, Will Be Handed To Community". UploadVR. https://www.uploadvr.com/mozilla-hubs-shutdown/. Retrieved 2026-10-04.
- ↑ Barrett Ens, Joel Lanir, Anthony Tang, Scott Bateman, Gun Lee, Thammathip Piumsomboon, Mark Billinghurst (2019-11). "Revisiting collaboration through mixed reality: The evolution of groupware". International Journal of Human-Computer Studies, vol. 131, pp. 81-98. Elsevier. doi:10.1016/j.ijhcs.2019.05.011. https://doi.org/10.1016/j.ijhcs.2019.05.011. Retrieved 2026-10-04.