In this episode, Filippo Sanfilippo is joined by Claudio Pacchierotti, a CNRS researcher at IRISA and an expert in haptics, wearable interfaces, robotic teleoperation, and shared control. They explore how touch can become a communication channel between humans and machines, and why haptic feedback matters for safer, more intuitive human-robot teaming.We discuss how autonomy can shift fluidly between humans and robots depending on task difficulty, confidence, and environmental uncertainty, plus where wearable haptics are heading next and which applications stand to benefit most.

Key topics

• Filippo Sanfilippo introduces human-robot teaming as the next step beyond human-robot interaction and collaboration, where humans and robots share space, forces, and decisions.
• Claudio explains haptics as a communication channel that adds a physical layer to human-machine interaction, making control more natural and safe.
• We discuss why touch is still underexplored compared with vision and audio, even though it becomes critical when humans and robots physically interact.
• Claudio describes shared control and shifting autonomy, where the robot takes on more or less of the task depending on confidence, task difficulty, and sensor quality.
• The episode looks at how haptic feedback supports situational awareness without overloading the visual channel.
• Claudio explains why wearable haptics matter for teleoperation and training, especially when natural dexterity must be preserved.
• We discuss cutaneous haptics such as vibration, skin stretch, pressure, and temperature, and how these can be combined with vision and audio.
• Claudio highlights practical application areas including drone fleets, humanoid robots, hospitals, logistics, medicine, space, defense, delivery, and factories.
• The conversation closes on the idea of human-centered robotics, where robots empower humans rather than replace them.

Timestamps

00:00 - Introducing Claudio Pacchierotti and the idea of human-robot teaming

00:32 - Claudio’s background in haptics, wearable interfaces, and shared control

01:31 - Why touch matters in human-machine interaction

03:05 - From human-robot collaboration to shared space, forces, and decisions

04:09 - Haptics as a communication channel between humans and robots

06:26 - Shared control for complex systems like humanoid robots and drone fleets

08:26 - How haptic feedback improves situational awareness during control

09:56 - Why touch can convey physical and abstract information without visual overload

11:15 - Shifting autonomy based on task difficulty and confidence

13:10 - Using shifting autonomy for training and complex robotic systems

15:17 - Cutaneous haptics, safe teleoperation, and keeping dexterity intact

17:22 - Combining vibration, pressure, skin stretch, and temperature

18:48 - Where human-robot teaming matters most in real-world applications

21:05 - Better actuators, sensors, algorithms, and haptic interfaces for teaming

23:13 - Why rich feedback improves trust, acceptance, and collaboration

24:56 - “Just enough haptics” and the future of wearable interfaces

27:12 - Human-centered robotics that empower rather than replace people

30:07 - Why the field is broad, multidisciplinary, and still expanding

Notable quotes

“Haptics as a way of communication between humans and cyber-physical systems.”

“We believe in robots that can empower humans, not replace humans.”

“Just enough haptics.”

Human-Robot Teaming, Randomness, and Why Statistics Matter

In this episode, Filippo Sanfilippo speaks with Professor Svein Olav Nyberg from the University of Agder about how statistics and probability sit at the center of human-robot teaming. They connect robotics to bachata dancing, human movement prediction, teaching, machine learning, and even dreams and consciousness.The conversation is wide-ranging, but the through line is clear: humans are not perfectly predictable, and good systems need to handle uncertainty rather than pretend it does not exist.

Key topics

• In this episode: Filippo introduces Svein Olav Nyberg as a colleague, researcher, author of Bayesian Way, and one of the statistical foundations behind the group’s human-robot teaming work.
• The discussion defines human-robot teaming (HRT) as more than interaction or collaboration: humans and machines share space, forces, and intention, and the leader-follower role can shift depending on the task.
• Bachata dancing becomes the metaphor for control design, especially the idea that one partner leads while the other follows, and that those roles can sometimes be swapped.
• The hosts explain a Greek-letter autonomy parameter, alpha, which acts like a slider for delegating more or less control from human to robot.
• A major example is trajectory prediction in industrial settings, where synthetic human motion is generated instead of forcing real workers to repeat paths thousands of times.
• Svein Olav describes using an Ornstein-Uhlenbeck process to model random motion with a pull back toward an intended path, including both position and velocity in two dimensions.
• The same movement data can support more than prediction: it can help slow robots near passing workers, adapt machine behavior to the workspace, identify operators by gait, and detect anomalies.
• Randomness becomes a recurring theme, from heart rate variability and walking patterns to art, teaching, and memory retention.
• Svein Olav shares how introducing randomness in lectures and exams can increase engagement and improve memory, because interest boosts retention.
• The conversation broadens to machine learning and statistics, with the point that statistics is the engine under the hood, especially when uncertainty matters more than a single output number.
• They discuss robotics moving from rigid white-box control toward hybrid systems that combine low-level motor and sensor control with statistical and machine learning methods.
• In the closing section, the guests reflect on future systems that can handle unexpected input, on pseudo-randomness in cryptography, and on whether machines might someday dream or feel emotions.

Timestamps

00:00 - Opening the episode and introducing the guest 01:21 - Why statistics matters for human-robot teaming 03:21 - From human-robot interaction to shared space, force, and intention 04:20 - Bachata as a metaphor for leader-follower control 05:47 - Swapping roles with AI during dance instruction 06:46 - The autonomy slider, alpha, in human-robot teaming 10:03 - Modeling human trajectories in industrial environments 11:00 - Generating synthetic humans instead of collecting exhausting real-world data 12:57 - Ornstein-Uhlenbeck motion modeling for position and velocity 14:23 - Using the same data for robot adaptation, identification, and anomaly detection 16:57 - Randomness, heart rate variability, and healthy human rhythms 19:42 - Random teaching, dice-based exams, and student engagement 21:29 - Why interest improves memory retention 23:30 - Statistics as the bridge between theory and practice in AI 24:56 - Why uncertainty matters more than a single machine learning output 27:24 - Bridging traditional robotics with statistics and machine learning 28:14 - Randomness in cryptography and pseudo-random number generators 30:26 - A football example that unexpectedly became a teaching tool 33:11 - Dreams, hallucinations, and whether machines can interpret or have dreams 34:49 - Blade Runner, consciousness, and synthetic beings 37:26 - The future is more open because systems must handle the unexpected 39:10 - Closing thoughts, karaoke, and the episode sign-off

Key frameworks

• Human-robot teaming
• Shared space
• Shared forces
• Shared intention
• Autonomy slider
• A controllable degree of delegation between human and robot
• Ornstein-Uhlenbeck process
• Random motion with a tendency to return toward a target state
• Statistics under the hood
• Machine learning is powerful, but statistics explains uncertainty and credibility

Notable quotes

"The future is far more open than we actually believe it is."

"Statistics is the engine" behind machine learning.

"We share space, we share forces, and we finally also share intention."

Join us for an enlightening discussion with PhD candidate Martin Økter about the forefront of human-robot teaming (HRT). We delve into how touch, machine learning, simulation, and safety are shaping the future of robotics across industries such as automotive recycling, healthcare, and space exploration.

Main Topics:

• The role of haptics and touch in enhancing human-robot interaction
• Imitation learning and simulation for robot training and safety
• Modular robotics architectures and ROS for complex applications
• Future implications and societal benefits of collaborative robots
• Practical insights for aspiring researchers and industry leaders

Timestamps:

00:00 - Introduction to human-robot teaming and Martin Økter

00:30 - The significance of haptics and the senses in robotics

01:02 - Applications of HRT in battery disassembly processes

01:33 - The shift from humans teaching robots to joint teamwork

02:02 - Using machine learning and imitation learning in robots

02:31 - The importance of tactile feedback and learning through testing

03:01 - Augmented surgical aid: reducing reliance on cadavers

03:30 - Designing scalable robotic systems with ROS

04:11 - Modular control architectures for complex disassembly tasks

04:49 - Simulation environments like Nvidia's IsaacSim for training robots

05:20 - Safety constraints and merging traditional control with machine learning

06:00 - Challenges and opportunities in long-term robot deployment

06:45 - Future scenarios: robotic applications in medical and industrial sectors

07:30 - The importance of continuous learning and exploration in robotics

08:10 - Advice for young researchers: passion, curiosity, and perseverance

08:50 - The societal role of collaborative robots and ethical considerations

09:40 - Concluding thoughts: a future driven by human-robot collaboration

Resources & Links:

• Nvidia IsaacSim - Nvidia's physics-based simulation platform
• Robot Operating System (ROS) - Modular framework for robot software development

Connect with Martin Økter:

• LinkedIn

Join us on this journey to the future of collaborative robotics, where humans and machines support each other, overcoming challenges and unlocking new potentials!