UbiCom Book Slides Chapter 6 Tagging, Sensing & Controlling Stefan Poslad http://www.eecs.qmul.ac.uk/people/stefan/ubicom Ubiquitous computing: smart devices, environments and interaction 1 Overview • • • • • • • • Introduction Tagging the Physical World Sensors and Sensor Networks Micro Sensing & MEMS Micro Actuation & MEMS Embedded Systems and Real-time Systems Control Systems (For Physical World Tasks) Robots Ubiquitous computing: smart devices, environments and interaction 2 Chapter 6: Overview The slides for this chapter are also expanded and split into several parts in the full pack • Part A: Tagging physical world & augmented reality • Part B: Sensors, Sensor Nets • Part C: MEMS • Part D: Embedded Systems • Part E: Control Systems & Robots Ubiquitous computing: smart devices, environments and interaction 3 Overview Chapter 6 focuses on: • internal system properties: context-awareness & autonomy • external interaction with the physical environment. Ubiquitous computing: smart devices, environments and interaction 4 Introduction To enable Smart (Physical) Environments, devices should: • Spread more into the physical environment, becoming part of more user activities in physical environment • Be cheap to operate: autonomous energy etc • Be low maintenance: automatic • Be able to interact with physical environment context • Be sometimes small enough so as to … • Be able to be encapsulated and embedded • Be cheap to manufacture Ubiquitous computing: smart devices, environments and interaction 5 UbiCom Internal System Properties Physical Environments Physical Phenomena CPI (Sense, CPI Adapt) implicit HCI Autonomous ContextAware Distributed Intelligent UbiComp System ICT Ubiquitous computing: smart devices, environments and interaction 6 Smart Physical Environments Nanotechnology Integrated Circuits Robots Embedded RT Nanobots MEMS Micro sensors actuators & Operating systems Process Control Macro Control Systems Sensor Nets Virtual Tags RFID Augmented Reality Locators Annotation Smart Physical Environments Ubiquitous computing: smart devices, environments and interaction 7 Smart (Physical) Environments Smart Devices CPI Context-aware systems Physical Environment Devices Natural Physical Environment OS Contexts Types Dimensions ASOS Controllers Tags MTOS Sensors macro micro Actuators Augmented Reality RTOS Sensor Nets Nano Types Physical Virtual Data Mgt MEMS Skins Programmable Paint Matter RFID Link PID Dust Robot Active Site, Anchor Etc. Passive Adaptive Arms Wheels Mobile Nanobot Legs Ubiquitous computing: smart devices, environments and interaction 8 Overview • • • • • • • Introduction Tagging the Physical World Sensors and Sensor Networks Micro Actuation and Sensing: MEMS Embedded Systems and Real-time Systems Control Systems (For Physical World Tasks) Robots Ubiquitous computing: smart devices, environments and interaction 9 Tagging (or Annotating) the Physical World Outline of this section • Applications • Life-cycle for Tagging Physical Objects • Tags: Types and Characteristics • Physical and Virtual Tag Management • RFID Tags • Personalised and Social Tags Ubiquitous computing: smart devices, environments and interaction 10 Tagging: Applications • Locate items, e.g.? • Retrieve annotations associated with physical objects (augmented reality) e.g. ? • Security, e/g/. . • Tracking, e.g., • Automated Routing: of physical objects, e.g., ? • Automated Physical Access: e.g., ? Ubiquitous computing: smart devices, environments and interaction 11 Tagging Applications: Automated Physical Access Ubiquitous computing: smart devices, environments and interaction 12 Tagging Applications: Asset Tracking Ubiquitous computing: smart devices, environments and interaction 13 Tagging Applications: Security Ubiquitous computing: smart devices, environments and interaction 14 Physical versus Virtual Tags Virtual View of physical objects, e.g., digital Photo Physical Tag, e.g., RFID Stefan’s car Virtual Tag Physical Object Ubiquitous computing: smart devices, environments and interaction 15 Life-cycle for Tagging Physical Objects Managing : Accessing : Presenting Capturing: Anchoring : . Organising: Ubiquitous computing: smart devices, environments and interaction 16 Design issues for Anchoring Tags on Physical Objects Different ways to characterise and classify tagging • By how to augment physical world objects for use in virtual (computer) environments • By use of Onsite versus Offsite and attached versus detached classification of tags Ubiquitous computing: smart devices, environments and interaction 17 Augment physical environments for use in virtual environments • Augment the user: • Augment the physical object: • Augment the surrounding environment: Ubiquitous computing: smart devices, environments and interaction 18 Onsite versus Offsite & Attached versus Detached Annotation 2 dimensions: • User of the annotation is – onsite (co-located or local) with physical object versus – offsite (not co-located or remote). • Annotation is – attached (or augments) physical object it refers to versus – being detached (not augmented or not collocated) with the physical object. Ubiquitous computing: smart devices, environments and interaction 19 Onsite versus Offsite & Attached versus Detached Annotation Attached Detached Offsite Onsite Ubiquitous computing: smart devices, environments and interaction 20 Design issues for Anchoring Tags on Physical Objects Physical Environment Objects Static States Dynamic States Tags Sensors Digital Analogue Physical Virtual (annotation) AR Physical-Virtual Tag Link Augment User RFID Onsite versus off-site Augment Physical Object Cardinality Augment Physical Environment Static vs. Dynamic Attached versus detached Ubiquitous computing: smart devices, environments and interaction Design issues for Tagging Physical environment • Tags read outdoors in noisy, wet, dark or bright environments. • Annotation data storage, distribution & integration with data • Data management must start as soon as the data is captured (readers). • Multiple tags & readers per unit Vol.. – Challenges? • Redundant annotations: similar items are captured, many times over. – Solutions? • Applications and businesses need to define the level of aggregation, reporting, analysis Ubiquitous computing: smart devices, environments and interaction 22 RFID Tags • A type of on-site tag, attached to physical object • RFID (Radio Frequency Identifier) Tags, attached to objects to enable identification of objects in the world over a wireless link. RFID Tags versus Bar codes? Ubiquitous computing: smart devices, environments and interaction 23 RFID Tags: Applications • ??? Ubiquitous computing: smart devices, environments and interaction 24 Types of RFID Tag • RFID tags may be classified into whether or not they: – Active: – Passive:. • Active tags are more expensive and require more maintenance but have a longer range compared to passive tags. • Typical RFID system main components: • tag itself, reader, data storage, post-processing • RFID tag versus RF Smart Card? Ubiquitous computing: smart devices, environments and interaction 25 Active RFID Tags • • • • • • Active RFID tags used on large, more expensive assets . Typically operate at 0.455, 2.45 or 5.8 GHz frequencies Have a read range of 20 M to 100 M, Cost? Complex active tags could also incorporate sensors. How? Why? • 2 types of active tags: transponders and beacons Ubiquitous computing: smart devices, environments and interaction 26 Active RFID Transponders • Active transponders are woken up when they receive a signal from a reader. • Transponders conserve battery life. How? • Important application of active transponders is in toll payment collection, checkpoint control and other systems. Ubiquitous computing: smart devices, environments and interaction 27 Active RFID Beacons • Main difference c.f. Transponder is long range, global? beacon reader • Beacons are used in Real-Time Location Systems (RTLS) • Longer range RTLS could utilise GPS or mobile phone GSM trilateration – See Chapter 7 • In RTLS, a beacon emits a signal with its unique identifier at pre-set intervals – Ubiquitous computing: smart devices, environments and interaction 28 Active RFID Transponder Application: toll booths Ubiquitous computing: smart devices, environments and interaction 29 Passive RFID Tags • • • • • Contain no power source and no active transmitter Power to transmit comes from where? Cheaper than active tags, cost? Shorter (read access) range than active tags, typically ?? Passive RFID transponder consists of a microchip attached to an antenna, e.g., same as smart card • Lower maintenance • Passive Transponders can be packaged in many different ways, – ???? Ubiquitous computing: smart devices, environments and interaction 30 Passive RFID Tags • Passive tags typically operate at lower frequencies than active tags – • Low-frequency tags are ideal for applications where the tag needs to be read through certain soft materials and water at a close range. Why? Ubiquitous computing: smart devices, environments and interaction 31 Passive Tags: Near Field • 2.different approaches to transfer power from the reader to passive tags: near field and far field Near field • Passive RFID interaction based upon electromagnetic induction. • Explain how this works here Ubiquitous computing: smart devices, environments and interaction 32 Passive Tags: Far Field • Why can’t electromagnetic induction be used? • So how does far field RFID interaction work? Ubiquitous computing: smart devices, environments and interaction 33 Business Use of Annotation • Physical artefact annotation is often driven by business goals. • Uniquely identify objects from manufacture during business processes Ubiquitous computing: smart devices, environments and interaction 34 Personal use of Annotation • Tags are less specific, deterministic, multi-modal (using multiple sensory channels) using multimedia. • Subjective annotations are used in multiple contexts, multiple applications and multiple activities by users. • Semantic gap challenge: between the low-level object features extracted and their high-level meaning with respect to a context of use • Several projects to tag personal views of physical world – – – – MyLifeBits Semacode Google Earth? But Is it personalised? etc Ubiquitous computing: smart devices, environments and interaction 35 Personal use of Annotation: Semacode • Semacode (2005) propose a scheme to define labels that can be automatically processed from captured images and linked to a Web-based spatial information encyclopaedia. • How does a semacode encodes URLs?? • How to create a semacodes? • How do read a Semacaode • Some management may be needed to control malicious removal, movement and attachment. Ubiquitous computing: smart devices, environments and interaction 36 Semacode Use Attach to physical world Convert URL to visual code Photograph (Read Code) get Web Page post Phone Ubiquitous computing: smart devices, environments and interaction 37 Overview • • • • • • • Introduction Tagging the Physical World Sensors and Sensor Networks Micro Actuation and Sensing: MEMS Embedded Systems and Real-time Systems Control Systems (For Physical World Tasks) Robots Ubiquitous computing: smart devices, environments and interaction 38 Sensors: introduction • Sensors are transducers that convert some physical phenomenon into an electrical signal • Wireless sensors: • Sensors can be networked – sensor nets Ubiquitous computing: smart devices, environments and interaction 39 Sensor Applications Give some examples of sensor use • Cars • Computers • Retail, logistics: • Household tasks • Buildings • Environment monitoring • Industrial sensing & diagnostics Ubiquitous computing: smart devices, environments and interaction 40 Sensors Types Sensors can be characterised according to: • Passive (tags) vs. active • Single sensors vs sensor arrays vs sensor nets • Read-only program vs. re-programmable Ubiquitous computing: smart devices, environments and interaction 41 Sensors versus Tags • ??? Ubiquitous computing: smart devices, environments and interaction 42 Distribution field of phenomena that can be detected measured Physical Phenomena S S Sensor net S S S Storage S S S Sensor net Sensor net S S S S S Sensors that detect event User S Access Node Sensors that notify access node Internet Ubiquitous computing: smart devices, environments and interaction 43 Sensor Nets • Main components of a typical sensor network system are networked sensors nodes serviced by sensor access node. • Slightly different but compatible view of a sensor network is to view sensors as being of three types of node): – common nodes – sink nodes – gateway (access) • In scenario given earlier, some sensors in the network can act as sink nodes within the network in addition to the access node. • Concepts of sensor node & sensor net can be ambiguous: – A sensor can act as a node in a network of sensors versus there is a special sensor network server often called a sensor (access) node Ubiquitous computing: smart devices, environments and interaction 44 Sensor Net: Functions • The main functions of sensor networks can be layered in a protocol stack according to: – physical network characteristics, – data network characteristics – data processing and sensor choreography • Use small network protocol stack for sensor nets. Why? • Other conceptual protocol layered stacks could also be used instead to model sensor operation, Ubiquitous computing: smart devices, environments and interaction 45 Sensor Net: Functions Data processing Collaborative processing Data storage Internetwork Sensor to Network Sensor Electronics Event definition & processing In-situ processing Routing Intra vs. inter node Sensor distribution & density DSP Data discovery Data uncertainty Addressing RF , Optical transmission characteristics Physical environment characteristics Power management Ubiquitous computing: smart devices, environments and interaction 46 Sensors: Electronics Processing Storage Sensor Transducer Analogue Filter Amplifier ADC DSP Transceiver Modulator Power Power management Battery Antenna Transmitter Switch Demodulator Receiver Sensor Net Design: Signal Detection & Processing Positioning & coverage of networks is important. Why? Ubiquitous computing: smart devices, environments and interaction 48 Sensor Net Design: Positioning & Coverage • Given: sensor field (either known sensor locations, or spatial density) – Where to add new nodes for max coverage? – How to move existing nodes for max coverage? • Can Control – Area coverage: – Detectability: – Node coverage: Ubiquitous computing: smart devices, environments and interaction 49 Sensor Net Design: Improved SNR Through Using Denser Sensor Nets • Sensor has finite range determined by base-line (floor) noise level • Denser sensor field improves detection of signal source within range. How? Ubiquitous computing: smart devices, environments and interaction 50 Overview • Overview: Sensor Net Components & Processes • Physical Network: Environment, Density & Transmission • Data Network: Addressing and Routing • Data Processing: Distributed Data Storage & Data Queries Ubiquitous computing: smart devices, environments and interaction 51 Senor Net Design: Sensor Data Routing • • • • • Networking sensors versus networking computers? Sensors form P2P network with a mesh topology network Sensors are massively distributed and work in real-time No universal routing protocols or central registry. Each node acts a router and application host. Ubiquitous computing: smart devices, environments and interaction 52 Sensor Routing • Make sensor address resolution efficient • Data centric routing, – Directed Diffusion – Flooding – Gossiping • Routing classification – Network structure: flat, hierarchical, hybrid – By interaction protocol Ubiquitous computing: smart devices, environments and interaction 53 Sensor Networks vs. Ad Hoc Networks ??? Ubiquitous computing: smart devices, environments and interaction 54 Sensor Net Topologies • ?? Ubiquitous computing: smart devices, environments and interaction 55 Senor Net Design: In-Network Processing • Why perform In-Network Processing? Sensor Node Sensors Ubiquitous computing: smart devices, environments and interaction 56 Sensor Net: Data Storage & Retrieval • What designs/ architectures can we use for sensor net data storage an retrieval? Ubiquitous computing: smart devices, environments and interaction 57 Sensor Database System • Characteristics of a Sensor Network: • Can existing database techniques be reused? Ubiquitous computing: smart devices, environments and interaction 58 Sensor Net: Technologies, Kits & Standards • • • • Sun Spot: Java Berkeley Motes: TinyOS, C SPINE (Signal Processing in Node Environment) OGC Standards: SensorML etc Ubiquitous computing: smart devices, environments and interaction 59 Overview • • • • • • • Introduction Tagging the Physical World Sensors and Sensor Networks Micro Actuation and Sensing: MEMS Embedded Systems and Real-time Systems Control Systems (For Physical World Tasks) Robots Ubiquitous computing: smart devices, environments and interaction 60 Micro Actuation and Sensing: MEMS • • • • • Fabrication Micro-Actuators Micro-Sensors Smart Surfaces, Skin, Paint, Matter and Dust Downsizing to Nanotechnology and Quantum Devices Ubiquitous computing: smart devices, environments and interaction 61 Trend: Miniaturisation • Electronic components become smaller, faster, cheaper to fabricate, lower power & lower maintenance, they can be more easily deployed on a massive and pervasive scale. • MicroElectro Mechanical Systems (MEMS) are based upon IC Chip design • Possibilities for miniaturization extend into all aspects of life, & potential for embedding computing & comms technology quite literally everywhere is becoming a reality. • IT as an invisible component in everyone's surroundings • Extending the Internet deep into the physical environment Ubiquitous computing: smart devices, environments and interaction 62 Trend: IC Transistor Density • Gordon Moore (1965), Intel co-founder made a prediction, now popularly known as Moore's Law, which states that the number of transistors on an IC chip doubles ~ every 2 y • Does it mean that software processing capability will also increases in this way? • IC Chip density = Software Performance? Ubiquitous computing: smart devices, environments and interaction 63 MEMS: Introduction • MEMS (Micro-electromechanical systems): micron- to millimetre-scale electronic devices fabricated as discrete devices or in large arrays • MEMS perform 2 basic types of functions: sensors or actuators. • Both act as transducers converting one signal into another. • MEMS actuators: electrical signal -> physical phenomena to move or control mechanisms. • MEMS Sensors work in reverse to actuators Ubiquitous computing: smart devices, environments and interaction 64 MEMS Examples Actuator Gyroscope Hinge Electrostatic motor Ubiquitous computing: smart devices, environments and interaction 65 MEMS: Fabrication • MEMS comprising mechanical and discrete electronic components • MEMS design is different from macro devices • MEMS design are based upon IC chips design • Silicon based materials have: – Well understood electrical properties – Good mechanical properties Ubiquitous computing: smart devices, environments and interaction 66 MEMS: Fabrication • Design a new circuit = design of interconnections among millions of relatively simple and identical components. • Diversity and complexity of the interconnections -> diversity of electronic components including memory chips and CPUs. • Multiplicity, batch fabrication, is inherent. • Miniaturisation of IC based MEMS processing has important advantages over macro electromechanical devices and systems? Ubiquitous computing: smart devices, environments and interaction 67 MEMS : Fabrication • Micromachines are fabricated just like ICs. • MEMS type ICs can be fabricated in different ways using: – Bulk micro-machining – Surface micro-machining – LIGA deep structures. Ubiquitous computing: smart devices, environments and interaction 68 Micro-Actuator • Mechanisms involved in micro-actuation whilst conceptually similar to equivalent macro mechanisms may function fundamentally differently, • Are engineered in a fundamentally different way using IC Ubiquitous computing: smart devices, environments and interaction 69 Micro-actuator: Applications • • • • • Micro-mirrors, e.g., ?? Micro-fluid pumps, e.g., ?? Miniature RF transceivers, e.g., ?? Miniature Storage devices, e.g., ?? Etc Ubiquitous computing: smart devices, environments and interaction 70 Micro-sensors • Sensors are a type of transducer • Microsensors can work quite differently from equivalent macro sensor, • Sensors enable adaptation • Often embedded into system as part of a control loop Ubiquitous computing: smart devices, environments and interaction 71 MEMS: Applications • Micro-accelerometers, – E.g., ?? • Micro-gyroscopes – E.g., • Detecting Structural Changes – E.g., Ubiquitous computing: smart devices, environments and interaction 72 Smart Device Form Factors: Smart Dust, Skins & Clay • 3 forms proposed by Weiser (1 tabs, 2 pads & 3 boards) can be extended to include 3 more forms: 4. Smart Dust: 5. Smart Skins: 6. Smart Clay: Ubiquitous computing: smart devices, environments and interaction 73 Smart Dust: MEMS • MEMS can be sprayed into physical environment • E.g., Smart Dust project (Pister, UC,Berkely) • (see Chapter 2) Ubiquitous computing: smart devices, environments and interaction 74 Smart Skins: MEMS • MEMS can be permanently attached to some fixed substrate forming – smart surfaces – smart skin • E.g. Paint that is able to sense vibrations • See also Organic Displays (Chapter 5) Ubiquitous computing: smart devices, environments and interaction 75 Smart Clay: MEMS • • • • • Claytronics project Can behave as malleable programmable matter Are MEMS ensembles Self-assembled into any arbitrary 3D shape Goal to achieve a synthetic reality. Ubiquitous computing: smart devices, environments and interaction 76 MEMS: Challenges • Establishing ownership of all of these micro items. • Coping with data overload • Different Low-level patterns of signals may be ambiguous and variable. • Handling context switches between these augmented environment events via assisted senses and the unassisted ones. • Are micro-devices either easy to dispose of or hard to dispose of? – What is we swallow / breath them in? • How to manage MEMS? – See Chapter 12 Ubiquitous computing: smart devices, environments and interaction 77 Nanocomputing • Nanocomputing can be defined as the manipulation, precision placement, measurement, modelling, and manufacture to create systems with less than 100 nm • Also referred to as nanotechnology • Is based upon a broader range of materials, mechanisms & sizes down to molecular level • MEMS Vs. Nanocomputing? Ubiquitous computing: smart devices, environments and interaction 78 Nanocomputing • The drive to switch transistors faster and to be lowpowered has been to make them smaller. • When electronic components approach nanometer sizes, odd things begin to happen. What? • This raised an early concern about the feasibility of nanotechnology. Other challenges are: • thermal noise • positioning and the control of structures at this level Ubiquitous computing: smart devices, environments and interaction 79 Nanocomputing • Nanotechnology at first proposed to use a bottom-up approach to design, to be able to assemble custom-made molecular structures for specific applications, • A major challenge to this design process is the complexity and novelty in understanding and being able to model materials at this level. • More research is needed to understand how combinations of materials, in particular compounds, gives materials at the molecular level certain physical and functional properties.. Ubiquitous computing: smart devices, environments and interaction 80 Overview • • • • • • • Introduction Tagging the Physical World Sensors and Sensor Networks Micro Actuation and Sensing: MEMS Embedded Systems and Real-time Systems Control Systems (For Physical World Tasks) Robots Ubiquitous computing: smart devices, environments and interaction 81 Embedded Systems: Introduction • • • • Is a component in a larger system Is programmable Performs a single, dedicated task. May or may not be visible as a computer to a user of that system • May or may not have a visible control interface • E.g., ??? • May be local or remote, – e.g., ?? • fixed or mobile – e.g?? Ubiquitous computing: smart devices, environments and interaction 82 Embedded System Characteristics (Embedded vs. MTOS Systems) Traditionally, embedded systems differ from MTOS systems OS of Embedded systems differ vs. MTOS system 1. Specialised to single task enactment (ASOS) 2. Actions on physical world tasks are often scheduled with respect to real-time constraints (RTOS) 3. Safety-criticality is considered more important Ubiquitous computing: smart devices, environments and interaction 83 Embedded vs. MTOS Systems • Often have constraints concerning power consumption • Often are designed to operate over a wide-range of physical environmental conditions compared to PC – e.g., • Often operate under moderate to severe real-time constraints. • System failures can have life-threatening consequences. – E.g., Ubiquitous computing: smart devices, environments and interaction 84 Embedded vs. MTOS Systems • Each embedded computing devices may be designed for its own rigidly defined operational bounds – e.g., • • • • • • Linking embedded systems to external systems Designs often engineered for a trade-off Fewer system resources then PC. How? Embedded systems not always easy to programme. Why? Most embedded designs (hardware & software) are unique Use a far simpler & cheaper OS & hardware. Why? Ubiquitous computing: smart devices, environments and interaction 85 Embedded Systems: Hardware • Microprocessors • Microcontroller • FPGA (Field Programmable Gate Arrays): Ubiquitous computing: smart devices, environments and interaction 86 Real-Time System (RTS) • Real-time systems (RTS) can be considered to be resource-constrained • Often RTS perform safety-critical tasks • RTS reacts to external events that interrupt it: • RTS uses mechanisms for priority scheduling of interrupts • RTOS may also use additional process control: – . Ubiquitous computing: smart devices, environments and interaction 87 RTS Design Concerns • There are a range of real-time design concerns to support critical response time of a task: – • Need to optimise – both response time and data transfer rate – optimising these when there are simultaneous tasks. • Key factors that affect the response time are? – process context-switching – interrupt latency Ubiquitous computing: smart devices, environments and interaction 88 RTS: Hard vs. Soft • Timeliness is single most important aspect of RT system. • RTS system is one where timing of result is just as important as the result itself. • A correct answer produced too late is just as bad as an incorrect answer or no answer at all. • RTS correctness of computations not only depends upon the logical correctness of the computation but also upon time to produce results. • If the timing constraints are not met, system failure occurs • Timing constraints can vary between different real-time systems. • Therefore, RTS can fall into one of three categories: soft, hard or firm.. Ubiquitous computing: smart devices, environments and interaction 89 RTS: Soft • Single computation arriving late may not be significant to the operation of the system, – • Although many late arrivals might be significant • Timing requirements can be defined by using an average response time. Ubiquitous computing: smart devices, environments and interaction 90 RTOS: Hard • • • • Timing requirements are vital. Response that’s late is incorrect and system failure results. Activities must complete by specified deadline, always. Different types of deadlines. What? • If a deadline is missed the task fails – E.g., ?? • This demands that the system has the ability to predict how long computations will take in advance. Ubiquitous computing: smart devices, environments and interaction 91 Safety-Critical Systems • Instructors could add some text here or delete this slide. Ubiquitous computing: smart devices, environments and interaction 92 Overview • • • • • • • Introduction Tagging the Physical World Sensors and Sensor Networks Micro Actuation and Sensing: MEMS Embedded Systems and Real-time Systems Control Systems (For Physical World Tasks) Robots Ubiquitous computing: smart devices, environments and interaction 93 Links to other Topics • Control systems / robots can be simple, operate in static deterministic environments. • To operate in more dynamic non- deterministic environments, they can make use of AI techniques (Chapters 8-10). • HCI aspects of (biologically inspired) robots such as affective computing etc (Chapter 5) Ubiquitous computing: smart devices, environments and interaction 94 Control Systems (For Physical World Tasks) • Simply type of control – Activated only when defined thresholds are crossed, – e.g., . • Disadvantages? – • Solutions? – Ubiquitous computing: smart devices, environments and interaction 95 Control Systems: Feedback Control • 2 basic kinds of feedback: – negative – positive Negative feedback • Seeks to reduce some change in a system output or state • Based upon derivative of output • Which is then used to modify input to regulate output. • Several types of feedback control: D, P, I, PID Positive feedback • Acts to amplify a system state or output Ubiquitous computing: smart devices, environments and interaction 96 Control Systems: Derivative (D) Feedback Control Reference Value r(t) ∑ - Control System + Controller Error e(t)=r(t)–f(t) f(t) DAC & Drive Input i(t) ADC Plant Output o(t) Transducer Feedback Ubiquitous computing: smart devices, environments and interaction 97 Control Systems: Proportional (P) Feedback Control • In simple proportional (-ve feedback) control system • Action taken to negatively feedback a signal to the plant, • Is in proportion to the degree the system diverges from the reference value • This leads to a much smoother regulation – e.g.,. P Controller e(t) Proportional g.e(t) Ubiquitous computing: smart devices, environments and interaction 98 Control Systems: PID Controllers • Sometimes P type controller output is not regulated correctly – e.g., ?? • To solve this problem either integral or differential control or both can be added to the control. • PID controller is so named because it combines Proportional, Integral and Derivative type control • Proportional (P) controller is just the error signal multiplied by a constant and fed out to a hardware drive. Ubiquitous computing: smart devices, environments and interaction 99 Control Systems: PID Controllers • Integral (I) controller deals with past behaviour of control. – • Derivative (D) type controller is used to predict the plant behaviour • P, PI, PD or PID control are often simple enough, to be hard-coded into controllers • Usually support some adjustment controls, – e.g., • PID controllers can be designed to be programmable Ubiquitous computing: smart devices, environments and interaction 100 PID Controllers PID Controller Proportional + Integral e(t) f(t) + ∑ Derivative Ubiquitous computing: smart devices, environments and interaction 101 Programmable Controllers: Microcontrollers • Hardware architecture of microcontrollers is much simpler than general purpose processor mother-boards in PCs? • I/O control support can be simpler as there may not be any video screen output or keyboard input. • Micro-controllers can range in complexity • Originally, programmed in assembly language, later in C • Control programs often developed in an emulator on a PC • More recent microcontrollers can be integrated with on-chip debug circuitry accessed by an in-circuit emulator Ubiquitous computing: smart devices, environments and interaction 102 Complex Control Systems • PID control Useful for coarse-gained, static control – E.g., palletising, coarse-controlled locomotion, etc • PID control not suitable for ? – fine-grained – dynamic control – uncertainties in control Ubiquitous computing: smart devices, environments and interaction 103 Complex Control • Several sources of uncertainty? • Techniques for controlling uncertain systems? Ubiquitous computing: smart devices, environments and interaction 104 Overview • • • • • • • Introduction Tagging the Physical World Sensors and Sensor Networks Micro Actuation and Sensing: MEMS Embedded Systems and Real-time Systems Control Systems (For Physical World Tasks) Robots Ubiquitous computing: smart devices, environments and interaction 105 Robots • Early 1960s, robots started to be used to automate industrial tasks particularly in manufacturing Why Automate? Ubiquitous computing: smart devices, environments and interaction 106 Main Robot Components Robots consist of: • End effectors or actuators: • Locomotion: • Drive: • Controller • Sensor Ubiquitous computing: smart devices, environments and interaction 107 Robots: Localisation • Localisation is used to determine a robot’s position in relation to its physical environment. • Localisation can be local or global. • Local localisation is often simpler in which a robot corrects its position in relation to its initial or other current reference location. • Global localisation is discussed more in context-aware systems part. Ubiquitous computing: smart devices, environments and interaction 108 Robots: Types 3 Main Types • Robot manipulator or robot arm • Mobile robots • Biologically inspired robots Ubiquitous computing: smart devices, environments and interaction 109 Robot Manipulators • A manipulator consists of a linked chain of rigid bodies that are linked in an open kinematic chain at joints. • rigid body can have up to 6 Degrees Of Freedom (DOF) of movement. • This comprises 3 translational DOF – ??? Ubiquitous computing: smart devices, environments and interaction 110 Robot Manipulators • Also comprises 3 rotational DOF – ??? • Joints are designed to restrict some DOF. • Human operators may be in the control loop of robot manipulators. Why? Ubiquitous computing: smart devices, environments and interaction 111 Robot Manipulators: Design • • • • Motion planning needed Control algorithms? Regulation of contact force Manipulators need to cope with variations in components and objects being manipulated. Solutions? – Use adaptive AI techniques (Chapter 8) – Put human in the control loop Ubiquitous computing: smart devices, environments and interaction 112 Mobile Robots • Mobile robots use various kinds of locomotion systems – ? • Simplest types of mobile robots to control – ?? • In dynamic non-deterministic environments, control is more complicated – • A more complex, well-known & highly successful use of mobile robots was Mars Explorer Robots – Ubiquitous computing: smart devices, environments and interaction 113 Mobile Robots • No. of DOF is often less compared to a robot manipulator. – Need ways to navigate obstacles? • Simple approach: use collision detection • More complex approach: anticipate & avoid collisions – Need environment models (AI, Chapter 8) – Need to replan paths to reach goal destinations (AI, Chapter 8) Ubiquitous computing: smart devices, environments and interaction 114 Biologically Inspired Robots: Legged locomotion • Biologically inspired robots are more complex type of robot – Combines legged locomotion capabilities & manipulator • 2 main focuses to these robots: – Legged locomotion (in combination with manipulator) – Human-Robot Interaction Ubiquitous computing: smart devices, environments and interaction 115 Biologically Inspired Robots: Legged locomotion • The use of legs enables legged robots to travel over irregular terrain • Biped robots often have more DOF than either the mobile robot or robot manipulator • Particular design challenge for biped robots is stability – Ubiquitous computing: smart devices, environments and interaction 116 Biologically Inspired Robots: Human Robot Interaction • Human robot Interaction: – a specialisation of HCI, see Chapter 5 • Robots can assist humans and extend sensing capabilities of (less able?) humans – Posthuman model. • Robots can fulfil social roles – i.e., affective computing (Chapter 5) – e.g., artificial pets • Social guided learning – Learning by imitation or by tutelage • Use of more human oriented interface & interaction – E.g., speech recognition Ubiquitous computing: smart devices, environments and interaction 117 Nanobots • Nanobots can be manufactured as MEMS or at molecular level. • Microscopic world is governed by the same physical laws as the macroscopic world • But relative importance of the physical laws change in how it affects the mechanics and the electronics at this scale Ubiquitous computing: smart devices, environments and interaction 118 Nanobots • Nature in terms of micro-organisms can be harnessed in order to provide a host body for nanobots to move about – e.g., • Shrinking device size to these nano dimensions leads to many interesting challenges: Ubiquitous computing: smart devices, environments and interaction 119 Developing UbiCom Robot Applications • Industrial types of robots • Low cost consumer type robots • Robots toolkits that are programmable. Ubiquitous computing: smart devices, environments and interaction 120 Light Sensor Motor A Ultrasonic Sensor Motor B Motor C Ubiquitous computing: smart devices, environments and interaction 121 Developing UbiCom Robot Applications • Task: robot manipulates a Rubik’s Cube to its solved state • Goal: robot performs whole task or guides humans to do it Design involves • Design: of the robot mechanics – • Design: how and when the robot senses state of the world – e.g. , • Planning algorithm: to link individual actions • Overall architecture: to integrate different sub-tasks – e.g., Ubiquitous computing: smart devices, environments and interaction 122 Developing UbiCom Robot Applications Several practical issues for physical robots tasks execution • Sensor accuracy • Position accuracy • Variable amounts of friction during movement • Some elasticity in the robot arm • • • • Low-level design to tell robots to carry out specific tasks Tasks need to be designed to fit the robots capabilities In open physical world, much non-determinism to handle -> There does not yet exist, flexible general purpose UbiCom robots, which can act as autonomous assistants or servants for mass human use. Ubiquitous computing: smart devices, environments and interaction 123 Overview • • • • • • • Introduction Tagging the Physical World Sensors and Sensor Networks Micro Actuation and Sensing: MEMS Embedded Systems and Real-time Systems Control Systems (For Physical World Tasks) Robots Ubiquitous computing: smart devices, environments and interaction 124 Summary & Revision For each chapter • See book web-site for chapter summaries, references, resources etc. • Identify new terms & concepts • Apply new terms and concepts: define, use in old and new situations & problems • Debate problems, challenges and solutions • See Chapter exercises on web-site Ubiquitous computing: smart devices, environments and interaction 125 Exercises: Define New Concepts • Annotation Ubiquitous computing: smart devices, environments and interaction 126 Exercise: Applying New Concepts Ubiquitous computing: smart devices, environments and interaction 127 Supplementary Slides • Exercises & Solutions Ubiquitous computing: smart devices, environments and interaction 128 Sensor Applications Ex: Give some examples of sensor use • Cars: air pressure, brake-wear, car-doors, engine etc • Lap-top: accelerometers – switch off computer disks when dropped • Retail, logistics: RFIDs • Heaters: thermostats • Infrastructure protection / Intrusion detection (active sensors) • Environment monitoring • Industrial sensing & diagnostics • Battlefield awareness • Sensors can be characterised according to: – passive (tags) vs. active – Single sensors vs sensor arrays vs sensor nets – Read-only program vs. re-programmable Ubiquitous computing: smart devices, environments and interaction 129