Detailed slide-by-slide text content extracted from this presentation.
Slide 01
Robotics Engineering
- Machines That Move, Sense, and Think
- From Ancient Automata to Autonomous Machines
- A 3,000-year journey from clockwork wonders to AI-powered systems reshaping every industry
- 1 / 32
Slide 02
What Is Robotics Engineering?
- Robotics engineering is the interdisciplinary field concerned with the design, construction, operation, and application of robots — machines capable of sensing their environment, processing information, and acting on the physical world.
- Core Disciplines
- Mechanical Engineering: Structure, actuators, end-effectors
- Electrical Engineering: Sensors, motors, power systems
- Computer Science: Software, AI, machine learning
- Control Theory: Feedback loops, stability, PID control
- The Sense-Think-Act Cycle
- Every robot operates on a fundamental loop:
- Sense: Perceive the environment (cameras, LiDAR, force sensors)
- Think: Process data, plan actions (algorithms, AI)
- Act: Execute movements (motors, actuators, end-effectors)
- "A robot is a machine that senses, thinks, and acts."
- — Rodney Brooks, former director of MIT CSAIL, co-founder of iRobot and Rethink Robotics
- 2 / 32
Slide 03
Ancient and Early Automata
- The dream of building artificial beings is older than recorded history. Ancient engineers created astonishing mechanical devices centuries before the word "robot" existed.
- ~400 BCE
- Archytas of Tarentum builds a steam-powered wooden pigeon that allegedly flew 200 meters — the first known autonomous machine. Described by Aulus Gellius in the 2nd century CE.
- ~250 BCE
- Ctesibius of Alexandria invents the water clock (clepsydra) with automated figures, and programmable cam-driven mechanisms — forerunners of industrial automation.
- 1206 CE
- Al-Jazari publishes The Book of Knowledge of Ingenious Mechanical Devices, describing 100+ automata including a programmable humanoid band that played music on a boat, and the first known crankshaft mechanism.
- 1495
- Leonardo da Vinci designs an "Automaton Knight" — an armored figure driven by pulleys, cables, and gears that could sit, stand, raise its visor, and wave its arms. Reconstructed in 2002, it worked exactly as designed.
- 3 / 32
Slide 04
The Birth of the Word "Robot"
- The word "robot" entered the world through theater, and the field's foundational ideas came from science fiction before engineering realized them.
- R.U.R. (1920)
- Czech playwright Karel Capek coined "robot" (from Czech robota, meaning forced labor) in his play Rossum's Universal Robots, premiered January 25, 1921 in Prague. The robots revolt and exterminate humanity — establishing the "robot uprising" narrative a century before AI anxiety.
- Asimov's Three Laws (1942)
- Isaac Asimov proposed three laws of robotics in his short story "Runaround":
- 1. A robot may not injure a human or allow harm through inaction
- 2. A robot must obey human orders (except when conflicting with Law 1)
- 3. A robot must protect its own existence (except when conflicting with Laws 1 or 2)
- These laws remain the starting point for modern AI ethics discussions, though Asimov himself wrote stories showing their inadequacy.
- 4 / 32
Slide 05
The First Industrial Robots
- 1954
- George Devol patents the first programmable robotic arm, the Unimate. His patent (#2,988,237) described a "Programmed Article Transfer" device capable of repeating stored movements.
- 1961
- The first Unimate is installed at a General Motors die-casting plant in Ewing, New Jersey. Weighing 4,000 lbs, it extracted hot metal parts from die-casting machines — a task too dangerous for human workers. Cost: $25,000 (equivalent to ~$260,000 today).
- 1969
- The Stanford Arm, developed by Victor Scheinman, becomes the first electrically powered, computer-controlled robotic arm. Its 6 degrees of freedom (DOF) set the standard for industrial manipulators.
- 1974
- ASEA (now ABB) introduces the IRB 6 — the first fully electric, microprocessor-controlled industrial robot. Sweden and Japan rapidly adopt industrial robotics; by 1980, Japan operates 14,000 industrial robots vs. 3,000 in the US.
- 5 / 32
Slide 06
Robot Anatomy: Mechanical Systems
- Actuators (Muscles)
- Electric motors: DC servo, stepper, brushless DC — most common in industrial and mobile robots
- Hydraulic: High force density, used in heavy construction (excavators, Boston Dynamics Atlas)
- Pneumatic: Compliant, lightweight, used in soft robotics and grippers
- Shape memory alloys: Nitinol wires contract when heated; used in micro-robots
- Artificial muscles: Electroactive polymers, dielectric elastomers (research stage)
- End-Effectors (Hands)
- Grippers: Parallel jaw, angular, vacuum suction, magnetic
- Tool changers: Automatic swap between welding torches, drills, cameras
- Dexterous hands: Shadow Dexterous Hand (24 DOF, 20 actuators, 129 sensors)
- Soft grippers: Silicone-based, compliant — can handle eggs, fruit, and irregular objects
- Degrees of freedom (DOF) define a robot's flexibility: a standard industrial arm has 6 DOF (3 position + 3 orientation). The human arm has 7 DOF; the human hand adds 27 more.
- 6 / 32
Slide 07
Sensing: How Robots Perceive the World
- Vision
- Cameras: RGB, stereo, depth (Intel RealSense, ZED 2). LiDAR: 3D point clouds, 150m range (Velodyne, Ouster). Used for SLAM (Simultaneous Localization and Mapping) — building a map while navigating it.
- Touch & Force
- Force/torque sensors: 6-axis measurement at the wrist. Tactile arrays: Pressure-sensitive skins (BioTac fingertip: 19 electrodes). Enable delicate manipulation — picking up a grape without crushing it.
- Proprioception
- Encoders: Measure joint angles (absolute/incremental). IMUs: Accelerometers + gyroscopes for orientation. Joint torque sensors: Measure internal forces. Critical for balance in legged robots.
- "Perception is the bottleneck. A robot that can see and feel as well as a 3-year-old child would revolutionize every industry."
- — Daniela Rus, Director of MIT CSAIL
- 7 / 32
Slide 08
Control Systems: The Robot's Brain
- Control theory is the mathematical framework that enables robots to execute precise, stable movements. Without it, a robot arm would oscillate wildly or collapse.
- PID Control
- The workhorse of robotics. A PID controller continuously adjusts motor output based on three terms:
- P (Proportional): Corrects based on current error
- I (Integral): Corrects based on accumulated past error
- D (Derivative): Corrects based on rate of error change
- 90%+ of industrial robot controllers use some form of PID.
- Advanced Control
- Model Predictive Control (MPC): Optimizes over a future time horizon. Used in Boston Dynamics' Atlas for dynamic balance
- Impedance Control: Regulates force and position simultaneously — essential for safe human-robot interaction
- Reinforcement Learning: Robot learns control policies through trial and error. DeepMind's work on robotic manipulation (2020–) achieves superhuman dexterity on specific tasks
- 8 / 32
Slide 09
Industrial Robotics Today
- 3.9M
- Industrial robots operating worldwide (2023)
- 553K
- New installations in 2023
- $16.5B
- Global industrial robot market (2023)
- Top Manufacturers
- CompanyCountrySpecialty
- FANUCJapanCNC + robotics, yellow arms
- ABBSwitzerlandHeavy industry, collaborative
- KUKAGermanyAutomotive, now Midea-owned
- YaskawaJapanWelding, servo motors
- Universal RobotsDenmarkCobots (collaborative robots)
- Robot Density (per 10K workers)
- South Korea: 1,012 (highest in the world)
- Singapore: 730
- Germany: 415
- Japan: 397
- China: 392 (grew from 49 in 2013)
- United States: 285
- World average: 151
- 9 / 32
Slide 10
Collaborative Robots (Cobots)
- Cobots are designed to work alongside humans without safety cages — a paradigm shift from traditional industrial robotics where robots operate in fenced-off cells.
- Universal Robots (UR)
- Founded 2005 in Odense, Denmark. Their UR5 (2008) was the first commercially successful cobot. Cobots now represent 10%+ of new industrial robot installations. Key features: force-limited joints (max 150N contact force), no-code programming via tablet, payload 3–20 kg, price $25K–$50K.
- Safety Standards
- ISO 15066 (2016): Defines force/pressure limits for human-robot contact
- Safety-rated monitored stop: Robot freezes when human enters workspace
- Speed and separation monitoring: Robot slows as human approaches
- Power and force limiting: Joints are compliant; impact force capped
- Hand guiding: Human physically moves robot to teach positions
- "The cobot revolution isn't about replacing humans. It's about making humans superhuman."
- — Esben Ostergaard, co-founder of Universal Robots
- 10 / 32
Slide 11
Mobile Robots: Wheels, Tracks & Legs
- Wheeled
- Simplest, most efficient on flat surfaces. Differential drive (Roomba), omnidirectional (Mecanum wheels), Ackermann steering (cars). Amazon's Kiva/Proteus warehouse robots move 800+ lbs at 5 mph using QR code floor navigation.
- Tracked
- Better traction on rough terrain. iRobot's PackBot (used in Afghanistan, Fukushima) weighs 24 lbs and can climb stairs. Tracked robots handle rubble, sand, and snow where wheels fail.
- Legged
- Most versatile on unstructured terrain. Boston Dynamics' Spot (2019): 4 legs, 14 kg payload, 5.2 km/h, 90-minute battery. Atlas (humanoid): backflips, parkour, dynamic balance. Agility Robotics' Digit: bipedal warehouse robot, being piloted at Amazon.
- Navigation: Modern mobile robots use SLAM (Simultaneous Localization and Mapping) algorithms that fuse LiDAR, cameras, and IMU data to build real-time 3D maps while navigating. Google Cartographer (2016, open-source) democratized SLAM for researchers.
- 11 / 32
Slide 12
Boston Dynamics: Pushing Physical Limits
- Founded in 1992 as a spin-off from MIT by Marc Raibert, Boston Dynamics has produced the world's most agile and dynamic robots.
- Key Robots
- BigDog (2005): DARPA-funded quadruped that could carry 340 lbs over rough terrain. Never deployed (too noisy)
- Atlas (2013–): Humanoid, 1.5m tall, 89 kg. Performs parkour, backflips, and object manipulation. Hydraulic (Gen 1) transitioning to electric (Gen 2, 2024)
- Spot (2019): Commercial quadruped, $74,500. Used in construction, oil & gas, public safety. 1,500+ units deployed globally
- Stretch (2021): Mobile warehouse robot, moves 800 boxes/hour
- Corporate Journey
- 2013: Acquired by Google (X) for ~$500M
- 2017: Sold to SoftBank for undisclosed amount
- 2020: Sold to Hyundai Motor Group for $1.1B (80% stake)
- Hyundai plans to integrate Spot and Atlas into its manufacturing and logistics operations
- 12 / 32
Slide 13
Autonomous Vehicles
- Self-driving vehicles are robots — they sense their environment, process complex real-time data, and act by steering, accelerating, and braking.
- SAE Autonomy Levels
- Level 0: No automation (manual driving)
- Level 1: Driver assistance (adaptive cruise control)
- Level 2: Partial automation (Tesla Autopilot, GM Super Cruise)
- Level 3: Conditional automation (Mercedes DRIVE PILOT, 2023 — first legal Level 3)
- Level 4: High automation (Waymo, Cruise, in geofenced areas)
- Level 5: Full automation (no steering wheel — not yet achieved)
- Waymo: The Leader
- Alphabet's Waymo operates 700+ driverless robotaxis in Phoenix, San Francisco, and Los Angeles. Over 100,000 paid rides per week (2024). Their vehicles have driven 20+ million autonomous miles on public roads. Technology stack: 29 cameras, 4 LiDAR units, 6 radar sensors, generating 1TB of data per hour per vehicle.
- 13 / 32
Slide 14
Drones: Robots That Fly
- Unmanned Aerial Vehicles (UAVs) represent one of the fastest-growing sectors in robotics, spanning consumer, commercial, and military applications.
- DJI: The Dominant Force
- Founded in 2006 by Frank Wang in Shenzhen, China, DJI controls ~70% of the global consumer/commercial drone market. The Phantom (2013) created the consumer drone category. The Mavic series (2016–) fits in a backpack. Annual revenue: ~$4.3B (2023). Their drones use GPS, visual odometry, and obstacle-avoidance AI.
- Commercial Applications
- Agriculture: Crop spraying, NDVI mapping (DJI Agras T40 covers 50 acres/hour)
- Delivery: Zipline delivers blood in Rwanda/Ghana (4,000+ deliveries/day)
- Infrastructure: Bridge/pipeline inspection replaces human climbers
- Film: Aerial cinematography at 1/100th of helicopter cost
- Emergency: Search and rescue with thermal cameras
- $45B
- Global drone market (2025 est.)
- 870K+
- FAA-registered drones (USA)
- 300K+
- Part 107 certified US pilots
- 14 / 32
Slide 15
Surgical Robotics
- Robotic surgery combines the precision of machines with the judgment of human surgeons, enabling minimally invasive procedures with sub-millimeter accuracy.
- The da Vinci System
- Made by Intuitive Surgical (founded 1995), the da Vinci is the dominant surgical robot:
- 8,600+ systems installed worldwide (2024)
- 12+ million procedures performed
- 4 robotic arms with 7 DOF each (exceeding the human wrist)
- 10x magnification 3D stereoscopic vision
- Cost: $1.5–2.5M per system + $2,000–3,000/procedure in consumables
- Beyond da Vinci
- Medtronic Hugo: Modular, lower cost competitor (FDA cleared 2024)
- CMR Surgical Versius: UK-made, portable robotic arms
- MAKO (Stryker): Orthopedic; 300,000+ joint replacements
- Neuralink: Robotic system inserts 1,024 electrodes into brain tissue with 50-micron precision
- "The robot doesn't perform surgery. The surgeon performs surgery through the robot. It's a tool of extraordinary precision."
- — Dr. Atul Gawande, surgeon and author
- 15 / 32
Slide 16
Humanoid Robots: The Race to Walk
- Humanoid robots — bipedal machines with human-like form — represent robotics' greatest engineering challenge and its most compelling vision.
- Major Humanoid Programs
- Honda ASIMO (2000–2022): Pioneered bipedal walking. Could run at 9 km/h. Retired after 22 years
- Boston Dynamics Atlas (2013–): Most agile humanoid. Electric version (2024) has unprecedented range of motion
- Tesla Optimus (2022–): $20K target price. Designed for factory tasks. Tesla aims for mass production
- Agility Digit (2019–): Bipedal, designed for logistics. Piloting at Amazon
- Figure 01/02 (2023–): $675M funding (Bezos, NVIDIA, Microsoft). OpenAI partnership for language integration
- Why Humanoid?
- Human-shaped robots can operate in environments designed for humans — stairs, doorways, tools, vehicles. They don't require infrastructure modifications. The argument against: human form is not optimized for most tasks. A wheeled robot is more efficient on flat ground; a snake robot is better in pipes. The counterargument: versatility in the human world requires human form.
- 16 / 32
Slide 17
AI & Machine Learning in Robotics
- The convergence of modern AI with robotics is creating machines that can learn, adapt, and generalize to new situations rather than following rigid programs.
- Key AI Approaches
- Computer Vision: Object detection (YOLO), semantic segmentation, pose estimation. CNNs process camera input in real time
- Reinforcement Learning: Robot learns through trial and error. OpenAI trained a robotic hand (Dactyl) to solve a Rubik's cube using simulation (2019)
- Sim-to-Real Transfer: Train in simulation (NVIDIA Isaac Sim), deploy on real hardware. Reduces training from years to hours
- Foundation Models: Large language models (GPT, Gemini) enable robots to understand natural-language commands and reason about tasks
- Google DeepMind RT-2 (2023)
- A "Robotic Transformer" that combines a vision-language model with robotic control. RT-2 can interpret commands it has never seen before ("pick up the extinct animal" — it picks up the toy dinosaur). This represents a fundamental shift: robots that understand concepts, not just coordinates.
- "The biggest revolution in robotics won't come from better hardware. It will come from better algorithms."
- — Pieter Abbeel, UC Berkeley professor and co-founder of Covariant
- 17 / 32
Slide 18
Soft Robotics: Machines That Bend
- Soft robotics uses compliant, deformable materials instead of rigid links and joints, enabling safer human interaction and the ability to handle delicate objects.
- Materials & Actuation
- Silicone elastomers: Moldable, stretchable, biocompatible
- Pneumatic networks (PneuNets): Air-filled channels that bend when pressurized
- Dielectric elastomer actuators: "Artificial muscles" that contract with voltage
- Shape memory polymers: Change shape with temperature
- Hydrogels: Water-based actuators for biomedical use
- Applications
- Food handling: Soft Robotics Inc. grippers sort delicate produce without bruising
- Medical: Soft endoscopes that navigate intestinal curves
- Ocean exploration: Harvard's Octobot (2016) — the first entirely soft autonomous robot
- Wearable: Soft exosuits from Harvard's Biodesign Lab assist walking (15% energy reduction)
- 18 / 32
Slide 19
Space Robotics
- Robots have been humanity's advance scouts in space exploration, operating in environments no human could survive.
- Mars Rovers
- Sojourner (1997): First Mars rover. 11 kg, traveled 100 meters in 83 days
- Spirit & Opportunity (2004): Designed for 90 days; Opportunity lasted 14 years and drove 45 km
- Curiosity (2012): Car-sized (899 kg), nuclear-powered (plutonium RTG). Still operating after 12+ years, having climbed Mount Sharp
- Perseverance (2021): Collects and caches samples for future return. Deployed Ingenuity helicopter — the first powered flight on another planet (72 flights before retirement in 2024)
- Space Station Robotics
- Canadarm2 (2001): 17.6m robotic arm on ISS, 7 DOF, can move 116,000 kg payloads. Built by MDA Space (Canada)
- Dextre (2008): Two-armed robot that performs repairs astronauts would otherwise do during EVAs
- Robonaut 2 (2011): NASA's humanoid, designed to work alongside astronauts
- GITAI (2024): Japanese startup testing autonomous robot arms on ISS
- 19 / 32
Slide 20
Underwater Robotics
- Remotely Operated Vehicles (ROVs) and Autonomous Underwater Vehicles (AUVs) explore ocean depths that are too deep, dark, and pressurized for human divers.
- Key Systems
- Jason (WHOI): Deep-sea ROV, 6,500m depth rating. Explored the Titanic wreck and hydrothermal vents
- REMUS 600 (Kongsberg): AUV for ocean mapping, mine countermeasures. Used by 25+ navies
- SoFi (MIT, 2018): Soft robotic fish that swims alongside real fish without disturbing them
- Nereid Under Ice (WHOI): Hybrid ROV/AUV for under-ice exploration in the Arctic
- Applications
- Oil & gas: Pipeline inspection, subsea wellhead maintenance (Oceaneering, TechnipFMC)
- Science: Deep-sea biology, geology, oceanography
- Defense: Mine countermeasures, submarine rescue
- Archaeology: Wreck exploration (Titanic, ancient Mediterranean vessels)
- Aquaculture: Fish pen inspection, net cleaning (80,000+ units forecast by 2030)
- 20 / 32
Slide 21
Agricultural Robotics
- Farming faces a labor crisis: the global agricultural workforce is aging and shrinking while food demand grows. Robots are filling the gap.
- Current Technology
- Autonomous tractors: John Deere 8R (2023): fully autonomous plowing, seeding, and spraying using GPS + cameras. No driver required
- Precision spraying: Blue River Technology (acquired by Deere for $305M): AI identifies individual plants, reducing herbicide use by 90%
- Harvesting robots: Agrobot (strawberries), Abundant Robotics (apples, folded 2021). Picking delicate fruit remains one of robotics' hardest unsolved problems
- Weeding robots: FarmWise, Carbon Robotics (laser weeding at 200,000 weeds/hour)
- The Challenge
- Agriculture is robotics' hardest real-world domain: unstructured environments, variable lighting, deformable objects (plants), mud, rain, and the need for extreme gentleness (bruised fruit is unsaleable). A strawberry-picking robot must locate a ripe berry, assess its ripeness via color and size, approach without damaging adjacent fruit, and detach it cleanly — all in under 3 seconds.
- 21 / 32
Slide 22
Warehouse & Logistics Robots
- The e-commerce explosion has driven explosive growth in warehouse automation. Amazon alone operates 750,000+ robots across its fulfillment network.
- Amazon Robotics
- Acquired Kiva Systems in 2012 for $775M — the largest acquisition in Amazon's history at that time. Kiva's mobile robots carry shelves to human pickers. By 2024, Amazon has deployed 750,000+ units across 50+ fulfillment centers. Newer systems: Proteus (fully autonomous mobile robot), Sparrow (AI-powered pick-and-pack arm), and Digit (Agility's bipedal robot in pilot).
- Key Players
- Locus Robotics: Collaborative mobile robots for order fulfillment
- Geek+: Chinese AMR (Autonomous Mobile Robot) leader, 30,000+ units deployed
- Fetch Robotics (Zebra): Warehouse and data-center mobile robots
- AutoStore: Norwegian cube-storage system; robots operate on a grid above storage bins
- Covariant: AI-powered robotic picking for mixed SKU environments
- 750K+
- Amazon robots
- $18.3B
- Warehouse robotics market (2025 est.)
- 40%
- Faster order processing with AMRs
- 22 / 32
Slide 23
Swarm Robotics
- Inspired by social insects (ants, bees, termites), swarm robotics uses large numbers of simple robots that coordinate through local rules to achieve collective behaviors no individual robot could accomplish.
- Principles
- Decentralization: No leader; each robot follows identical simple rules
- Local communication: Robots interact only with immediate neighbors
- Emergent behavior: Complex group patterns arise from simple individual actions
- Redundancy: Any robot can fail without disrupting the swarm
- Scalability: Adding robots improves performance linearly
- Examples
- Harvard Kilobots (2014): 1,024 coin-sized robots self-organize into shapes
- Intel Shooting Star drones: 2,018 drones formed the Olympics rings at PyeongChang (2018)
- TERMES (Harvard): Construction robots that build structures like termites
- Swarm search and rescue: Distribute robots across collapsed buildings to find survivors
- 23 / 32
Slide 24
Exoskeletons & Wearable Robots
- Exoskeletons augment human strength, endurance, or mobility by attaching powered structures to the body.
- Medical Exoskeletons
- Ekso Bionics EksoNR: FDA-cleared for stroke and spinal cord injury rehabilitation. Used in 350+ rehab centers
- ReWalk: Enables paraplegic patients to stand and walk. FDA-cleared 2014. $77,000 per unit
- Cyberdyne HAL: Japanese exoskeleton that reads bioelectric signals from the skin to anticipate intended movements
- Industrial Exoskeletons
- Sarcos Guardian XO: Full-body powered exoskeleton; lifts 200 lbs repeatedly without strain
- Hilti EXO-O1: Passive overhead exoskeleton for construction workers; reduces shoulder fatigue by 47%
- German Bionic Cray X: Back-support exoskeleton; reduces spinal load by 30 kg per lift
- Hyundai VEX: Vest exoskeleton for automotive assembly workers
- 24 / 32
Slide 25
Robot Operating System (ROS)
- ROS is the de facto open-source software platform for robotics research and development — the "Linux of robotics."
- What ROS Provides
- Hardware abstraction layer (same code works on different robots)
- Inter-process communication via publish/subscribe messaging
- Package management (5,000+ community packages)
- Simulation via Gazebo (physics engine) and RViz (visualization)
- Libraries for SLAM, navigation, manipulation, computer vision
- History & Impact
- Created at Willow Garage (2007) by Morgan Quigley, ROS became open-source in 2008. ROS 2 (2017) added real-time support and security for commercial deployment. Used by NASA, Toyota Research, Fetch Robotics, and 80%+ of robotics research labs worldwide. ROS reduced the "time to first demo" for a new robot from months to days.
- "ROS didn't just change how we build robots. It changed how we think about robots — as software platforms, not hardware products."
- — Morgan Quigley, creator of ROS
- 25 / 32
Slide 26
Robot Ethics & Safety
- As robots become more autonomous and enter human spaces, fundamental ethical questions about responsibility, employment, and autonomy demand answers.
- Key Ethical Questions
- Liability: When a self-driving car kills a pedestrian, who is responsible? The programmer? The manufacturer? The "driver"?
- Employment: McKinsey estimates 400M–800M jobs could be displaced by automation by 2030. What obligations do deployers have?
- Lethal autonomy: Should robots ever make kill/no-kill decisions in warfare?
- Privacy: Domestic robots with cameras and microphones are surveillance devices
- Bias: If training data is biased, robot behavior inherits that bias
- Regulatory Landscape
- EU AI Act (2024): First comprehensive AI/robotics legislation. Classifies high-risk AI systems (medical robots, autonomous vehicles) for mandatory conformity assessment
- ISO 10218 / ISO 15066: Industrial and collaborative robot safety standards
- IEEE 7000 series: Standards for ethical AI and autonomous systems
- Campaign to Stop Killer Robots: 70+ nations support a ban on fully autonomous weapons
- 26 / 32
Slide 27
Military & Defense Robotics
- Military applications drive significant robotics funding and development, but also raise the most urgent ethical concerns about autonomous weapons.
- Current Systems
- MQ-9 Reaper: USAF drone, 27-hour endurance, $32M each. Remotely piloted from Nevada
- PackBot/TALON: 6,000+ ground robots deployed in Iraq/Afghanistan for bomb disposal
- MAARS: Armed ground robot with non-lethal and lethal options (human always in the loop)
- Ghost Robotics Vision 60: Armed quadruped for perimeter security (USAF trials)
- The LAWS Debate
- Lethal Autonomous Weapons Systems (LAWS) — weapons that can select and engage targets without human intervention — are the subject of intense UN debate since 2014. The concern: removing human judgment from kill decisions crosses a moral red line. As of 2024, no international treaty bans LAWS, though 30+ nations call for preemptive regulation.
- "The decision to take a human life should never be delegated to a machine."
- — International Committee of the Red Cross position statement (2021)
- 27 / 32
Slide 28
Robotics Competitions
- Competitions drive innovation by setting audacious challenges and inspiring the next generation of roboticists.
- Major Competitions
- DARPA Grand Challenge (2004/2005): Autonomous desert driving. The 2004 race: no vehicle finished. 2005: 5 vehicles completed the 132-mile course. Directly led to Google's self-driving car project
- DARPA Robotics Challenge (2015): Disaster-response robots. Tasks: drive a car, open a door, climb stairs. Many robots fell spectacularly — exposing the gap between lab demos and real-world performance
- RoboCup: Annual robot soccer tournament (since 1997). Goal: by 2050, a team of humanoid robots will beat the FIFA World Cup champions
- FIRST Robotics: 500,000+ high school students build competition robots annually
- XPRIZE Challenges
- ANA Avatar XPRIZE (2022): $10M for telepresence robots that convey a sense of physical presence
- XPRIZE Rainforest (2024): Autonomous exploration of tropical canopy using drones and ground robots
- These competitions routinely produce startups: iRobot, Boston Dynamics, and Waymo all trace lineage to DARPA challenges
- 28 / 32
Slide 29
The Robotics Market: By the Numbers
- $72B
- Global robotics market (2025 est.)
- 26%
- CAGR for service robots (2020–2025)
- $5.8B
- VC investment in robotics (2023)
- Market Segments
- Industrial: $16.5B (automotive, electronics, metals)
- Service (professional): $18B (logistics, medical, agriculture)
- Service (consumer): $9B (vacuum, lawn, companion)
- Military/defense: $15B+ (drones, ground robots, maritime)
- Collaborative: $2.4B (fastest-growing segment)
- Top Robotics Nations
- China: 52% of global industrial robot installations (2023)
- Japan: Largest robot manufacturer (FANUC, Yaskawa, Kawasaki)
- South Korea: Highest robot density per worker
- Germany: European leader (KUKA, Franka Emika)
- USA: AI and software leadership (Boston Dynamics, Waymo, NVIDIA)
- 29 / 32
Slide 30
Micro- and Nano-Robotics
- At the smallest scales, robots promise to revolutionize medicine by operating inside the human body at cellular resolution.
- Current Research
- Magnetic microrobots: Sub-millimeter devices steered by external magnetic fields. ETH Zurich's group has demonstrated targeted drug delivery in animal models
- DNA nanorobots: Wyss Institute (Harvard) built a DNA "cage" that opens to release drugs when it detects cancer markers (2012, published in Science)
- Sperm-driven microrobots: IFW Dresden attached magnetic caps to sperm cells, steering them to deliver drugs to tumors
- Wireless microelectrodes: Sub-mm devices for deep-brain stimulation without wired implants
- Challenges
- Power: No batteries at sub-mm scale; must use external fields (magnetic, acoustic, light)
- Communication: Cannot carry radio antennas; fluorescence or MRI imaging used for tracking
- Manufacturing: Photolithography, 3D laser printing, DNA origami
- Biocompatibility: Materials must not trigger immune response
- FDA approval pathway: No regulatory framework exists for autonomous nano-devices
- 30 / 32
Slide 31
The Future of Robotics
- 2025–2030 Predictions
- Humanoid robots enter commercial production (Tesla Optimus, Figure, 1X)
- Level 4 autonomous vehicles expand beyond geofenced areas
- Surgical robots become semi-autonomous for routine procedures
- Warehouse automation reaches 80%+ in major fulfillment centers
- Construction robotics tackles the global housing shortage
- 2030–2050 Horizon
- General-purpose humanoid robots in homes and eldercare
- Swarm robots for environmental restoration (ocean cleanup, reforestation)
- Medical nanorobots for targeted cancer therapy
- Lunar and Martian construction robots build habitats
- Human-robot cognitive collaboration — robots as thinking partners, not just tools
- "We are at the ChatGPT moment for robotics. Foundation models will give robots the ability to understand the world, not just navigate it."
- — Ken Goldberg, UC Berkeley, 2024
- 31 / 32
Slide 32
Building the Future, One Machine at a Time
- "The question is not whether robots will change the world. The question is whether we will design them wisely enough to change it for the better."
- — Cynthia Breazeal, MIT Media Lab, pioneer of social robotics
- From Al-Jazari's water-powered automata to NVIDIA-powered humanoids, robotics engineering is the discipline that gives machines the ability to act in the physical world. Its greatest challenges — dexterous manipulation, real-world perception, ethical deployment — are also its greatest opportunities.
- 32 slides · Robotics Engineering · 2024
- 32 / 32