top of page

Michael Faber

Establishing Core Standards for Autonomous Delivery Robots

Communication and Design Modalities Proposal for Autonomous Delivery Robot (ADR).

Role

UX Research

Methods / Tools

Double dimond

Benchmark research

Literature review

Interviews

Journey Mapping

Cliant / Framework

GM / Reichman University HCI MA Program Internship

Background

Project Motivation

Autonomous objects, especially in the field of transportation, are increasingly shaping our environments, from homes to public spaces. Traditionally solely focused on vehicle manufacturing, car companies are slowly changing their skin and starting to see themselves as robotics companies to come, for all intents and purposes. Those car manufacturers are vigorously expanding their involvement in developing autonomous mobile robots (AMRs) for various tasks, including autonomous delivery robots (ADRs) for package delivery. That, alongside other robotics startup companies and major retail players (such as Amazon), are all searching to lead the race.

This project was initiated as part of the General Motors R&D department and Reichman HCI MA program collaboration, aiming to contribute to the company's efforts to understand better the challenges this race brings.

Proxemics

Influence on proxemics behavior in Human-Robot Interaction (HRI)

In this part, I explored how physical proximity influences interactions and behaviors in Human-Robot Interaction (HRI), similar to human interactions. Understanding this "dance" when a robot meets a person is crucial, leading to both expected and surprising findings. Influential factors include:

Human Characteristics:

  • Personality: Extraverts are more tolerant of close proximity.

  • Age: Children maintain more distance than adults.

  • Previous Experience: Familiarity affects comfort levels.

  • Approaching Direction: Robot-approaching direction is influenced by gender preference.

Robot Characteristics:

  • Voice: Synthesized voices are less favored, and male tones tend to be preferred.

  • Form: Machine-like robots are favored over human-like ones.

  • Speed: Movements should be slower than human walking speed.

  • Height: Taller robots may be seen as more competent, but this effect varies.

In conclusion, putting it into actionable guidelines for proxemics:

The optimal robot design involves a height between 1.2 to 1.5 meters, a slower movement speed, and a non-human form. Proxemics guidelines include a learning algorithm to recognize people, a robot “head” for directional and communicative gestures without staring, and a human-like voice if used.

My research adopted a structured and multi-faceted approach. Beginning with established methodologies, I utilized the 8-step eHMI design process and the Double-Diamond methodology, tailoring them to the unique challenges in the field of AMRs. The following steps detail the comprehensive strategy employed in the research:​

  1. Spatial Dimension Study: Examining proxemics within Human-Robot Interaction (HRI) to understand spatial dynamics.

  2. Signifier Analysis: Analyzing communication cues and mental models from the automotive industry for potential application in AMRs.

  3. AMR Environment & Robot Encounters: Investigating how AMRs navigate and interact with their surroundings and other robots.

  4. AMR-Human Interactions: Assessing the nature of interactions between AMRs and humans and what can be learned from AMRs’ encounters with people so far.

  5. Industry Understanding: Researching various types of delivery, their characteristics, and insights from human-to-human delivery interactions.

  6. ADR Analysis: Review existing Autonomous Delivery Robots, their operational frameworks, market size, and conducting design audits.

  7. Focused Encounter Study: Diving deep into a specific interaction scenario using a state machine diagram.

This approach was designed to build a thorough understanding of AMRs and drive forward the creation of a comon AMR-Human langauge and the standardization of communication modalities.

Research Limitations

There are Always Some...

The research has encountered several limitations:

  • The scarcity of academic literature in this emerging field.

  • Challenges accessing operational robots for direct observation and testing.

  • A noticeable reluctance from commercial companies in the. courier and delivery sector to collaborate.

Also, the study's insights were somewhat geographically constrained, as it was conducted solely in Israel. Furthermore, the research duration was confined to the span of the internship framework and timeline, limiting the investigation's scope and depth.

The Spatial Dimension

Research Process

Applying the Double-Diamond Principle

Researching transitions of autonomous mobile robots (AMR) between different spatial dimensions

01

  • Proxemics

  • Communication & Signifiers

  • HRI, RRI, HRRI - encounters abstraction

Focusing on autonomous delivery robots (ADR)

02

  • Learning the industry (Interview with FedEx Marketing manager)

  • Macro user journey

  • Possible encounters

  • Benchmark & Design Audit

Focusing on ADR specific service (personal parcels delivery) and encounter

03

  • Micro user journey; State machine; Sequential diagram

  • User Interviews

  • Online resources and VLOGS

A proposal of modalities for the standardization of communication and design aspects for ADRs’

04

  • Future research (Prototype, WOZ, Pilot, etc.)

Overview

Problem Statement

While the automotive vehicle industry adheres to strict regulations and a joint body was established to define a standard protocol for an autonomous vehicle, similar frameworks for non-human transporting AMRs and ADRs as part of them are lacking. This project steps into this gap, aiming to establish guidelines for developing autonomous moving robots, particularly Autonomous Delivery Robots (ADRs).

Early in my research, a predominant issue emerged while learning the AMR transition in between spaces. Humans excel in communication through verbal cues, gestures, body language, and eye contact, but how does this translate to interactions with autonomous entities like AMRs? The core challenge is understanding the intentions of an Autonomous Delivery Robot (ADR): "Is it approaching me?"; "Will it avoid collision or should I be the one to step aside?" The central problem is the absence of a shared language between humans and ADRs, which becomes even more complex in multiple robot interactions. This gap in communication highlights the need for underscoring the need for clear interaction protocols.

Signifiers

Establish Communication Baselines

Signifiers can be divided into Motion-inherent cues, Non-Verbal explicit cues, and signaling methods. In comparison to the world of autonomous vehicles, signs and cues slightly shift when we talk about delivery robots. For instance, while cars move at certain speeds, a delivery robot usually strolls slower than our walking pace. In this case, I will not pretend to present which signal is the most correct but highlight where and what signals are needed, dividing them into spatial and conceptual signals. The exact type of these signals, though, is something we'll need to explore further.

Motion-inherent cues

  • Speed

  • Distance

  • Deceleration/ Acceleration

  • Braking

Non-Verbal explicit cues

  • Eye contact

  • Head nodes

  • Gesture

Signaling methods

  • Headlights

  • Signaling/ hazard lights

  • Horn

Learning the industry

AMRs’ Encounters

Environment, Other Robots and People

Looking into how robots move in space, we can categorize their encounters into three main types: with the environment (like entering buildings or crossing roads), with other robots (such as navigating swarms on sidewalks), and with people (like picking up a delivery). These examples illustrate the situations robots may face in their daily operations.

What is "Delivery"?

Benchmark, User Journey and Design Audit

Delivery service, which we encounter almost daily, has more to it than we might think at first. My initial step was to grasp its complexity: understanding the fundamentals of "delivery," the various types and their characteristics, closely observing user journeys in ordering deliveries, and identifying pain points in delivery services. Also, I explored the emerging world of robotic deliveries, examining existing delivery robots, their design, functionality, and market trends in implementing these services.

Shipment types vary widely, including factors like content (food, products, legal documents), physical characteristics (package size, weight, temperature needs), and shipping logistics (speed and distance). These factors influence the delivery method and any special courier requirements. For my study, I've chosen to focus on what is known as “express deliveries” as well as “Same day delivery” or “last mile delivery.” This type of service handles small to medium packages up to 50 kg within a 200 km radius and is suitable for motorcycle or car transport. It's important to note that food deliveries, a distinct category, are not covered in this study.

The Journey

Of Both the Robot and the User

To better understand the process from both the service provider - the robot, and the customer's perspective; I developed a macro mapping detailing their journey up to their point of interaction.

Market Behavior

Soaring Momentum

Though ADR technology acceptance is debated, and regulatory hurdles limit their use in public spaces, market analysts forecast a surge in delivery robot value to around $29B by 2029. That explains the market momentum, where giants like Amazon and FedEx have already launched robot delivery services, and what highlights the urgent need for a communication protocol that will guide robot design, aiming to enhance user experience and human-robot interaction.

Design Audit

Issues to Consider

As I examined the various types of operating ADRs, I have listed several issues to consider:

  1. Some are below the line of sight of pedestrians and car drivers (flag).

  2. Some require physical effort to bend over (elderly).

  3. Shared compartment - lack of privacy.

  4. Footprint per delivery (delivering air).

  5. Visual pollution in space.

  6. Competing with other surroundings signals.

User Interviews

What I Learned from the Current Human-Human Delivery Interaction

To understand what makes a good delivery experience, I turned to interviews with people aged 20-65, who regularly receive deliveries (by human currier) living in both high-rise buildings and private homes, in cities and less central areas.

These interviews covered everything from their expectations and experiences meeting delivery personnel to issues of trust, courier behavior, and concerns about privacy and safety. Paradoxically, the insights gathered hinted that robot delivery services might offer a more considerate and human-like service experience.

HRI design considerations:

Interviews Insights

  1. Opportunity for personalization of the service (proxemics - “The Dance”).

  2. Cultural based interaction design.

  3. Establish familiarity to enhance attachment in the human-robot interaction (giving a robot a name - “Hi Siri”).

  4. Prolonging the interaction. Provide a feeling that the robot is "thinking of me."

  5. Establish standardizations for eHMI communication protocol

  6. Service design is crucial.

What makes for a good experience for me is a good communication and coordination of the hand-off time, much before the courier arrives.

It's gotten to the point where I'm already exchanging a few words with him here and there because he knows me and tells me “so we meet again...", by now I know him by his name.

When I got here I had a cultural shock because in Argentina the courier doesn't go up to your apartment, you meet the courier downstairs, that's the norm.

The ADR should be designed to medium in size and height (1.2-1.5 and 50-70kg), divided into several compartments. Also, the design process should consider the ADRs’ personality aspect through form, voice, and movement. This configuration guidelines may provide:

  1. Optimal utilization of space.

  2. Variety in service (parcel types).

  3. Compatibility for various people with different needs.

  4. Increase acceptance.

Design Insights

Currently, out of the models that can be found in the market, Alibaba's' Cainiao robot (on the right) answers the design guidelines the closest way.

Focusing on a specific encounter

Personal Same-Day Delivery State Machine

At this point, my attention turned to mapping out the "dance," creating a state-machine diagram of the crucial moment of interaction between the robot and the recipient. Paying attention to the progression of emotions, thoughts, and actions that unfold during this encounter, considering the recipient's varying mindsets throughout the experience.

Setting Communication protocol for ADRs’

A Need for Tailored Approach

While applying the eHMI Communication Protocol from Autonomous Vehicles (AVs) to Autonomous Delivery Robots (ADRs) might seem feasible, it's important to recognize their distinct differences. Although AVs and ADRs share some basic rules, the way people perceive them —considering factors like perceived safety and their operational traits (e.g., speed) and the absence of a "driver" in ADRs— varies significantly.

Furthermore, the communication protocols designed for AVs require expansion to address the unique functional and practical differences inherent to ADRs.

Conclusively, adapting and enhancing the AV communication protocol for ADRs is crucial, underscoring the need for a tailored approach to ensure effective and safe interactions between ADRs and humans in the evolving landscape. This necessity highlights the importance of our work and paves the way for pioneering safer, more intuitive human-robot interactions.

Takeaways and What next

A Work In Progress

Uncovering the real challenge beneath the surface noise of apparent issues is a significant and key takeaway from this project. Identifying the core problem was crucial not just for devising the right solution but also for engaging stakeholders and facilitating effective communication among all project participants.

Working in a dynamic, cutting-edge company in the automotive industry and robot world has been a dream spot for me as a researcher and user experience specialist. It's where I got to lay down some early ideas for how future designs and tech might shape up—a really exciting point that's both challenging and super interesting.

Though we've made significant progress, there's still plenty ahead. My next steps involve developing a prototype at Reichman University's Milab lab for initial testing. This will help refine our communication models and design guidelines for delivery robots, aiming for practical applications in ongoing company projects.

Stay tuned for further updates.

bottom of page