That question matters for buyers because "companion robot" usually means something that sits on a table, rolls indoors, or starts a video call. A walking companion for seniors is a different category. It has to share sidewalks with people, deal with weather and noise, understand fatigue and hesitation, keep enough distance to feel respectful, and still be useful when a caregiver is not standing nearby.
The current home robot market already has pieces of that puzzle. ElliQ 3 shows what proactive older-adult companionship can look like indoors. temi V3 shows mature indoor navigation and telepresence. Mirokai shows a social mobile humanoid built for human spaces. iCub and ergoCub show why research platforms still matter before companies promise useful, mobile caregiving robots.
What Stanford SOAR Is Really Testing
SOAR is not just a product teaser. Stanford describes the program as work to improve the lives of older adults and caregivers through robotics, with a focus on movement, time outdoors, social connection, and cognitive health. The Stanford Robotics Center also places SOAR among its flagship projects, alongside work in domestic robotics, human-robot interaction, medical robotics, field robotics, and other areas where robots must operate around real people.
That combination is the important part. A senior walking companion is not only a navigation problem. It is a human-robot interaction problem, a care workflow problem, and a reliability problem. If the robot encourages someone to take a short walk, notices when the person slows down, or helps a remote caregiver understand whether the outing went well, it starts to provide value beyond novelty.
It also raises the bar. A robot that fails in a living room is annoying. A robot that fails during a walk can strand the user, block a path, create anxiety, or give a caregiver false confidence.
The Minimum Useful Feature Set
A practical walking companion robot for seniors needs more than autonomous movement. These are the features that should be treated as table stakes before the category is ready for ordinary homes:
Requirement
Outdoor navigation
- Why it matters at home
- Sidewalks, curbs, ramps, shadows, cyclists, pets, and parked cars are harder than clean indoor floors.
Requirement
Human pacing
- Why it matters at home
- The robot must adapt to slow walking, rest breaks, hesitation, and changes in gait without nagging.
Requirement
Social interaction
- Why it matters at home
- It should speak clearly, listen in noisy streets, and keep conversation optional rather than intrusive.
Requirement
Caregiver visibility
- Why it matters at home
- Families need simple summaries, check-ins, and escalation paths, not raw robotics telemetry.
Requirement
Battery honesty
- Why it matters at home
- A walking robot must know whether it has enough charge to go out, come back, and handle delays.
Requirement
Weather limits
- Why it matters at home
- Rain, heat, cold, glare, and uneven pavement should be explicit limits, not hidden fine print.
Requirement
Privacy controls
- Why it matters at home
- Cameras, microphones, location history, and caregiver access need plain-language settings.
| Requirement | Why it matters at home |
|---|---|
| Outdoor navigation | Sidewalks, curbs, ramps, shadows, cyclists, pets, and parked cars are harder than clean indoor floors. |
| Human pacing | The robot must adapt to slow walking, rest breaks, hesitation, and changes in gait without nagging. |
| Social interaction | It should speak clearly, listen in noisy streets, and keep conversation optional rather than intrusive. |
| Caregiver visibility | Families need simple summaries, check-ins, and escalation paths, not raw robotics telemetry. |
| Battery honesty | A walking robot must know whether it has enough charge to go out, come back, and handle delays. |
| Weather limits | Rain, heat, cold, glare, and uneven pavement should be explicit limits, not hidden fine print. |
| Privacy controls | Cameras, microphones, location history, and caregiver access need plain-language settings. |
The most important word here is "honesty." A walking companion does not have to solve every mobility case. It does need to communicate what it can and cannot do before someone trusts it with a routine.
How This Differs From Today's Companion Robots
ElliQ 3 is one of the clearest examples of an older-adult companion robot already aimed at the home. In ui44's database, ElliQ 3 is a 14.4-inch, 5.3-pound, mains-powered tabletop system with proactive conversation, wellness programs, reminders, video calling, and virtual activities. Its pricing model is a lease-plus-subscription structure: a $249 initiation fee with monthly plans listed at $59/month, $49/month annually, or $39/month on a 24-month commitment.
That is relevant because ElliQ is built around ongoing engagement, not physical support. It can prompt activity, offer social connection, and help with routines, but it does not accompany a person outside. For many older adults, that may be enough. For someone whose hardest part is leaving the apartment, crossing a lobby, or staying motivated during a walk, the missing body matters.
Mirokai is closer to the embodied social side. ui44 tracks it as a 123 cm, roughly 26 kg social humanoid from Enchanted Tools with an omnidirectional rolling globe base, expressive projected face, torque-controlled arms, autonomous navigation, multi-LLM conversational AI, and about four hours of battery life. It is not sold like a consumer gadget; Enchanted Tools uses a partnership and inquiry flow rather than public list pricing.
The lesson is that a believable walking companion may look less like a smart speaker and more like a service robot: mobile, expressive, aware of nearby people, and designed to share semi-public spaces. But Mirokai also shows the cost and deployment gap. Robots that can safely move through hospitals, hotels, and care settings are still closer to institutional pilots than impulse home purchases.
Telepresence Is Not Enough
temi V3 is useful to compare because it already solves many indoor mobility problems. ui44 lists temi V3 as a roughly 100 cm, 12 kg personal robot assistant with autonomous navigation, telepresence, voice interaction, a 13.3-inch touchscreen, 360-degree LiDAR, depth cameras, and up to eight hours of battery life. That is a strong indoor foundation.
For senior care, telepresence is valuable. A family member can appear on the screen, a clinician can check in, and the robot can move between rooms. But a walking companion must do something harder: it has to coordinate with the user's body in the real world. A remote video call can encourage a walk, but it cannot easily manage sidewalk pace, road crossings, crowding, glare, wind noise, or the social awkwardness of a robot trailing someone too closely.
This is why Stanford's outdoor framing is important. The buyer question is not "Can this robot navigate?" It is "Can this robot make a walk safer, easier, or more likely to happen?"
Research Platforms Still Matter
iCub is not a home product, but it explains why labs keep using expensive humanoids before assistive robots become consumer devices. ui44 lists iCub as a 104 cm, 22 kg open-source humanoid research platform with 53 degrees of freedom, stereo vision, microphones, optional tactile skin, and an indicative non-profit build fee of about €250,000 for a full unit. The Stanford SOAR page also lists cognitive training protocols with iCub as one of its collaborations.
That price makes iCub irrelevant as a consumer purchase, but the research value is clear. Older-adult robotics needs careful study of attention, trust, timing, tone, cognitive load, and physical ability. Those are not specs you can infer from motor torque alone.
ergoCub points in another direction: human-scale assistance. ui44 tracks ergoCub as a 150 cm, 55.7 kg humanoid research platform focused on ergonomic collaboration, load handling, worker-intention recognition, navigation, and manipulation. Its listed collaborative load capability is around 10 kg. That is not a walking companion either, but it shows the direction assistive robotics may eventually take: robots that understand human motion and help with physical tasks around people.
What Should Buyers Ask Before Believing The Category?
If walking companion robots start appearing in pilot programs, care communities, or premium home trials, buyers should ask practical questions first:
- Where has it been tested: lab, senior community, sidewalk, campus, apartment building, or private home?
- What happens if the user stops, turns back, sits down, or becomes confused?
- Does the robot help in an emergency, or only notify someone else?
- Can caregivers see useful summaries without constant surveillance?
- What surfaces, slopes, rain conditions, temperatures, and lighting does the robot refuse?
- How does it behave around strangers, dogs, children, cyclists, and narrow paths?
- Is the robot useful indoors too, or is it a single-purpose walking device?
- Is pricing subscription-based, leased, institutional, or a direct purchase?
The strongest products will answer these questions in normal language. If a vendor can only talk about AI, autonomy, and companionship in general terms, the robot is probably not ready for a high-trust care role.
Where The First Real Deployments May Happen
The first serious walking companions are unlikely to be mass-market home devices. They are more likely to show up in senior living communities, retirement campuses, hospitals, rehab centers, or university-backed pilots. Those environments offer repeatable routes, trained staff, charging locations, and a way to observe whether the robot improves activity rather than merely attracting attention.
That does not make the category less important. It makes it more credible. Senior care robotics should earn trust in controlled real-world environments before it asks families to trust it in public sidewalks and private routines.
For ui44 readers, the right comparison set is not a robot vacuum. It is the overlap between companion robots, telepresence robots, social service robots, and humanoid research systems. ElliQ shows the engagement layer. temi shows mobile indoor presence. Mirokai shows embodied social interaction in public-facing spaces. iCub and ergoCub show the research path toward more capable assistive bodies.
Bottom Line
Stanford SOAR is a sign that older-adult robotics is moving beyond tabletop companionship and into the harder question of how robots can support movement, outdoor life, and social connection. That is a better direction than another generic gadget promising "AI care" without specifying what it actually does.
A walking companion robot for seniors should not be judged by whether it looks friendly. It should be judged by whether it helps someone move more confidently, gives caregivers useful context, respects privacy, and knows when not to go outside. Until those basics are proven, the category belongs in pilots and research programs. Once they are proven, it could become one of the more meaningful home robot categories we track.
Sources: Stanford SOAR, Stanford Robotics Center, and ui44 robot database entries for ElliQ 3, Mirokai, temi V3, iCub, and ergoCub.
Written by
ui44 Team
Published July 4, 2026
Share this article
Open a plain share link on X or Bluesky. No embeds, no widgets, no cookie baggage.