Learning and Control in Assistive Robotics for the Elderly

Q. Meng and M. H. Lee, Learning and Control in Assistive Robotics for the Elderly, IEEE Conference on Robotics, Automation and Mechatronics (RAM), Singapore, 2004.

An Autonomous Sailing Robot for Ocean Observation

Sauze, C. and Neal, M. 'An Autonomous Sailing Robot for Ocean Observation', in proceedings of TAROS 2006, Guildford, UK, Sept 4-6th 2006, pages 190-197.

Developmental Learning for Autonomous Robots

M.H. Lee, Q. Meng and F. Chao, 'Developmental Learning for Autonomous Robots', Robotics and Autonomous Systems, 55(9), pp 750-759, 2007.

A Customizable Camera-based Human Computer Interaction System Allowing People With Disabilities Autonomous Hands Free Navigation of Multiple Computing Task

Many people suffer from conditions that lead to deterioration of motor control and makes access to the computer using traditional input devices difficult. In particular, they may loose control of hand movement to the extent that the standard mouse cannot be used as a pointing device. Most current alternatives use markers or specialized hardware to track and translate a user's movement to pointer movement. These approaches may be perceived as intrusive, for example, wearable devices. Camera-based assistive systems that use visual tracking of features on the user's body often require cumbersome manual adjustment. This paper introduces an enhanced computer vision based strategy where features, for example on a user's face, viewed through an inexpensive USB camera, are tracked and translated to pointer movement. The main contributions of this paper are (1) enhancing a video based interface with a mechanism for mapping feature movement to pointer movement, which allows users to navigate to all areas of the screen even with very limited physical movement, and (2) providing a customizable, hierarchical navigation framework for human computer interaction (HCI). This framework provides effective use of the vision-based interface system for accessing multiple applications in an autonomous setting. Experiments with several users show the effectiveness of the mapping strategy and its usage within the application framework as a practical tool for desktop users with disabilities.

SOVEREIGN: An Autonomous Neural System for Incrementally Learning Planned Action Sequence to Navigate Towards a Rewarded Goal

How do reactive and planned behaviors interact in real time? How are sequences of such behaviors released at appropriate times during autonomous navigation to realize valued goals? Controllers for both animals and mobile robots, or animats, need reactive mechanisms for exploration, and learned plans to reach goal objects once an environment becomes familiar. The SOVEREIGN (Self-Organizing, Vision, Expectation, Recognition, Emotion, Intelligent, Goaloriented Navigation) animat model embodies these capabilities, and is tested in a 3D virtual reality environment. SOVEREIGN includes several interacting subsystems which model complementary properties of cortical What and Where processing streams and which clarify similarities between mechanisms for navigation and arm movement control. As the animat explores an environment, visual inputs are processed by networks that are sensitive to visual form and motion in the What and Where streams, respectively. Position-invariant and sizeinvariant recognition categories are learned by real-time incremental learning in the What stream. Estimates of target position relative to the animat are computed in the Where stream, and can activate approach movements toward the target. Motion cues from animat locomotion can elicit head-orienting movements to bring a new target into view. Approach and orienting movements are alternately performed during animat navigation. Cumulative estimates of each movement are derived from interacting proprioceptive and visual cues. Movement sequences are stored within a motor working memory. Sequences of visual categories are stored in a sensory working memory. These working memories trigger learning of sensory and motor sequence categories, or plans, which together control planned movements. Predictively effective chunk combinations are selectively enhanced via reinforcement learning when the animat is rewarded. Selected planning chunks effect a gradual transition from variable reactive exploratory movements to efficient goal-oriented planned movement sequences. Volitional signals gate interactions between model subsystems and the release of overt behaviors. The model can control different motor sequences under different motivational states and learns more efficient sequences to rewarded goals as exploration proceeds.

Vector Associative Maps: Unsupervised Real-time Error-based Learning and Control of Movement Trajectories

This article describes neural network models for adaptive control of arm movement trajectories during visually guided reaching and, more generally, a framework for unsupervised real-time error-based learning. The models clarify how a child, or untrained robot, can learn to reach for objects that it sees. Piaget has provided basic insights with his concept of a circular reaction: As an infant makes internally generated movements of its hand, the eyes automatically follow this motion. A transformation is learned between the visual representation of hand position and the motor representation of hand position. Learning of this transformation eventually enables the child to accurately reach for visually detected targets. Grossberg and Kuperstein have shown how the eye movement system can use visual error signals to correct movement parameters via cerebellar learning. Here it is shown how endogenously generated arm movements lead to adaptive tuning of arm control parameters. These movements also activate the target position representations that are used to learn the visuo-motor transformation that controls visually guided reaching. The AVITE model presented here is an adaptive neural circuit based on the Vector Integration to Endpoint (VITE) model for arm and speech trajectory generation of Bullock and Grossberg. In the VITE model, a Target Position Command (TPC) represents the location of the desired target. The Present Position Command (PPC) encodes the present hand-arm configuration. The Difference Vector (DV) population continuously.computes the difference between the PPC and the TPC. A speed-controlling GO signal multiplies DV output. The PPC integrates the (DV)·(GO) product and generates an outflow command to the arm. Integration at the PPC continues at a rate dependent on GO signal size until the DV reaches zero, at which time the PPC equals the TPC. The AVITE model explains how self-consistent TPC and PPC coordinates are autonomously generated and learned. Learning of AVITE parameters is regulated by activation of a self-regulating Endogenous Random Generator (ERG) of training vectors. Each vector is integrated at the PPC, giving rise to a movement command. The generation of each vector induces a complementary postural phase during which ERG output stops and learning occurs. Then a new vector is generated and the cycle is repeated. This cyclic, biphasic behavior is controlled by a specialized gated dipole circuit. ERG output autonomously stops in such a way that, across trials, a broad sample of workspace target positions is generated. When the ERG shuts off, a modulator gate opens, copying the PPC into the TPC. Learning of a transformation from TPC to PPC occurs using the DV as an error signal that is zeroed due to learning. This learning scheme is called a Vector Associative Map, or VAM. The VAM model is a general-purpose device for autonomous real-time error-based learning and performance of associative maps. The DV stage serves the dual function of reading out new TPCs during performance and reading in new adaptive weights during learning, without a disruption of real-time operation. YAMs thus provide an on-line unsupervised alternative to the off-line properties of supervised error-correction learning algorithms. YAMs and VAM cascades for learning motor-to-motor and spatial-to-motor maps are described. YAM models and Adaptive Resonance Theory (ART) models exhibit complementary matching, learning, and performance properties that together provide a foundation for designing a total sensory-cognitive and cognitive-motor autonomous system.

Control of cyclin B1 localization through regulated binding of the nuclear export factor CRM1.

Activation of the Cyclin B/Cdc2 kinase complex triggers entry into mitosis in all eukaryotic cells. Cyclin B1 localization changes dramatically during the cell cycle, precipitously transiting from the cytoplasm to the nucleus at the beginning of mitosis. Presumably, this relocalization promotes the phosphorylation of nuclear targets critical for chromatin condensation and nuclear envelope breakdown. We show here that the previously characterized cytoplasmic retention sequence of Cyclin B1, responsible for its interphase cytoplasmic localization, is actually an autonomous nuclear export sequence, capable of directing nuclear export of a heterologous protein, and able to bind specifically to the recently identified export mediator, CRM1. We propose that the observed cytoplasmic localization of Cyclin B1 during interphase reflects the equilibrium between ongoing nuclear import and rapid CRM1-mediated export. In support of this hypothesis, we found that treatment of cells with leptomycin B, which disrupted Cyclin B1-CRM1 interactions, led to a marked nuclear accumulation of Cyclin B1. In mitosis, Cyclin B1 undergoes phosphorylation at several sites, a subset of which have been proposed to play a role in Cyclin B1 accumulation in the nucleus. Both CRM1 binding and the ability to direct nuclear export were affected by mutation of these phosphorylation sites; thus, we propose that Cyclin B1 phosphorylation at the G2/M transition prevents its interaction with CRM1, thereby reducing nuclear export and facilitating nuclear accumulation.

To grow or not to grow: nutritional control of development during Caenorhabditis elegans L1 arrest.

It is widely appreciated that larvae of the nematode Caenorhabditis elegans arrest development by forming dauer larvae in response to multiple unfavorable environmental conditions. C. elegans larvae can also reversibly arrest development earlier, during the first larval stage (L1), in response to starvation. "L1 arrest" (also known as "L1 diapause") occurs without morphological modification but is accompanied by increased stress resistance. Caloric restriction and periodic fasting can extend adult lifespan, and developmental models are critical to understanding how the animal is buffered from fluctuations in nutrient availability, impacting lifespan. L1 arrest provides an opportunity to study nutritional control of development. Given its relevance to aging, diabetes, obesity and cancer, interest in L1 arrest is increasing, and signaling pathways and gene regulatory mechanisms controlling arrest and recovery have been characterized. Insulin-like signaling is a critical regulator, and it is modified by and acts through microRNAs. DAF-18/PTEN, AMP-activated kinase and fatty acid biosynthesis are also involved. The nervous system, epidermis, and intestine contribute systemically to regulation of arrest, but cell-autonomous signaling likely contributes to regulation in the germline. A relatively small number of genes affecting starvation survival during L1 arrest are known, and many of them also affect adult lifespan, reflecting a common genetic basis ripe for exploration. mRNA expression is well characterized during arrest, recovery, and normal L1 development, providing a metazoan model for nutritional control of gene expression. In particular, post-recruitment regulation of RNA polymerase II is under nutritional control, potentially contributing to a rapid and coordinated response to feeding. The phenomenology of L1 arrest will be reviewed, as well as regulation of developmental arrest and starvation survival by various signaling pathways and gene regulatory mechanisms.

An allelic series reveals essential roles for FY in plant development in addition to flowering-time control.

The autonomous pathway functions to promote flowering in Arabidopsis by limiting the accumulation of the floral repressor FLOWERING LOCUS C (FLC). Within this pathway FCA is a plant-specific, nuclear RNA-binding protein, which interacts with FY, a highly conserved eukaryotic polyadenylation factor. FCA and FY function to control polyadenylation site choice during processing of the FCA transcript. Null mutations in the yeast FY homologue Pfs2p are lethal. This raises the question as to whether these essential RNA processing functions are conserved in plants. Characterisation of an allelic series of fy mutations reveals that null alleles are embryo lethal. Furthermore, silencing of FY, but not FCA, is deleterious to growth in Nicotiana. The late-flowering fy alleles are hypomorphic and indicate a requirement for both intact FY WD repeats and the C-terminal domain in repression of FLC. The FY C-terminal domain binds FCA and in vitro assays demonstrate a requirement for both C-terminal FY-PPLPP repeats during this interaction. The expression domain of FY supports its roles in essential and flowering-time functions. Hence, FY may mediate both regulated and constitutive RNA 3'-end processing.

Voltage Sensing Using an Asynchronous Charge-to-Digital Converter for Energy-Autonomous Environments

In future systems with relatively unreliable and unpredictable energy sources such as harvesters, the system power supply may become non-deterministic. For energy effective operations, Vdd is an important parameter in any meaningful system control mechanism. Reliable and accurate on-chip voltage sensors are therefore indispensible for the power and computation management of such systems. Existing voltage sensing methods are not suitable because they usually require a stable and known reference (voltage, current, time, frequency, etc.), which is difficult to obtain in this environment. This paper describes an autonomous reference-free voltage sensor designed using an asynchronous counter powered by the charge on a capacitor and a small controller. Unlike existing methods, the voltage information is directly generated as a digital code. The sensor, fabricated in the 180 nm technology node, was tested successfully through performing measurements over the voltage range from 1.8 V down to 0.8 V.

Discursive Control, Non-Domination and Hegelian Recognition Theory: Marrying Pettit’s Account(s) of Freedom with a Pippinian/Brandomian Reading of Hegelian Agency

The aim of this article is to combine Pettit’s account(s) of freedom, both his work on discursive control and on non-domination, with Pippin’s and Brandom’s reinterpretation of Hegelian rational agency and the role of recognition theory within it. The benefits of combining these two theories lie, as the article hopes to show, in three findings: first, re-examining Hegelian agency in the spirit of Brandom and Pippin in combination with Pettit’s views on freedom shows clearly why and in which way a Hegelian account of rational agency can ground an attractive socio-political account of freedom; second, the reconciling of discursive control and non-domination with Hegelian agency shows how the force and scope of recognition become finally tangible, without either falling into the trap of overburdening the concept, or merely reducing it to the idea of simple respect; third, the arguments from this article also highlight the importance of freedom as non-domination and how this notion is, indeed, as Pettit himself claims, an agency-freedom which aims at successfully securing the social, political, economic and even (some) psychological conditions for free and autonomous agency.