This application claims priority under 35 U.S.C. §119 to European Patent Application No. 12156837.2 filed on Feb. 24, 2013, the entire content of which is hereby incorporated by reference.
The present invention is related to wireless communication and particularly to the task of resource allocation for transmitters/receivers in a wireless network, where base stations are equipped with a plurality of antennas as used, for example, in MIMO scenarios.
The present invention is related to wireless communications, transmission technologies, proportional fairness allocations, opportunistic resource allocations and predictive scheduling.
The demand for higher data rates in wireless communications is ever increasing. Thus, one has to find ways to use the given resources even more efficiently. Gains can be achieved by exploiting temporal variations in the channels due to fading that is independent among the users, so-called multi-user diversity. Opportunistic resource allocation (scheduling) was introduced in [1]. Well recognized work in this field is [2] and [3]; an overview can be found in [4]. The drawback of these schemes, purely aiming at increasing throughput, is the unfairness and starvation of users. So one seeks a balance between maximizing throughput and having a fair resource allocation among the users.
Proportional fairness offers an attractive trade-off between resource efficiency by opportunistically exploiting time-variant channels and the satisfaction of the users. Proportional fair sharing (PFS) was introduced in [5, 6] for the Qualcomm High Data Rates system.
The PFS is designed for a single channel network with TDMA constraint, that is, only one user is allowed to transmit at the same time. An extension to a system with multiple channels, with equal power per carrier, is introduced in [7]. And a similar but less general approach specifically designed for the 3GPP LTE Uplink is [8]. In the following, systems that allow only a single user per resource block are called orthogonal access systems.
Further increase in spectral efficiency for future generation networks is established by advanced physical layer techniques, for example multi-user MIMO. In multi-user systems with adaptive modulation and coding, the data rates of the users are coupled and in theory infinitely many rate configurations can be provided. These systems are referred to as advanced multi-user systems. The complex interdependence of the user rates is a significant difference and unfortunately there is no straight-forward extension of the PFS rule to advanced multi-user systems. A step to design opportunistic and fair resource allocation for multi-user systems is the formulation as an optimization problem; for proportional fairness this is the maximization of the sum of logarithmic average user rates [9]. For the PFS algorithm the interpretation as utility maximization and proof for asymptotic optimality can be found in [10]. To formulate the utility maximization some assumptions and definitions are introduced to describe the system model.
System Model: Slotted time-varying wireless channels are assumed, where the channel is assumed to be static within one time-slot. The channel state H is a random process and H[T] is the channel state realization at time-slot T. A peak power constraint is assumed, which implies that power budgets cannot be exchanged among the time-slots, as for an average power constraint. Depending on the capabilities of the hardware, the set of achievable data rates for the set of users K, K=|K| at time-slot T are given by the rate region R(H[T])=R[T]. The instantaneous rates established in time-slot T are r[T] ∈ R[T]. The weighted sample mean of the data rates is
The weights can be used to establish various definitions of the average throughput, see
in case the weights and the stochastic process of the channel states are such that the limit exists. This allows to define a region of long-term average rate regions supported by the physical layer:
With these definitions one can state opportunistic and fair resource allocation as maximizing a utility of the long-term average throughput:
where the utility associated with proportional fairness is U(
The optimal long-term average throughput
This means: one cannot optimize the average throughput directly. Instead, one decides for a rate allocation r[t] in each time step, which then automatically results in a certain average throughput.
The goal is to find a close to optimal causal scheduling strategy for any time-slot t which only utilizes information about previously made decisions and previous channel state information which defines the rate regions. Under certain conditions the following policies are asymptotically (T→∞) optimal:
Gradient Method [11, 12] The rate configuration for the current time-slot t is based on maximizing a linear approximation of the utility:
For the proportional fairness utility we have
which leads to the well known PFS rule [5, 6] in case of a TDMA constraint, where a single user needs to be selected. Therefore the gradient method can be considered as a generalization of proportional fair sharing for orthogonal access systems to proportional fair resource allocation for advanced multi-user systems.
Stochastic Subgradient Method Another attempt to solve problem (1.1) causally is the stochastic subgradient method. The rate configuration of the current time-slot is
where λ[t] are the dual variables updated as follows
with a fixed constant α.
Methods from Queuing Theory The task of optimizing a network utility is also considered in the area of queuing networks [13, 14] and virtual queues can be used for allocating resources in a way that leads to an optimal solution with respect to the network utility.
The rate configuration for the current time-slot is
where u[t] is the virtual queue updated as follows
with a fixed constant β.
The work in [15] is mentioned that specifically treats multi-user MIMO, but does consider an average power constraint and can therefore not be applied to the present scenario without major modifications.
The algorithms are memoryless, in the sense that they do not require keeping track of the rate allocations in the past or channel states. Instead, they track a single variable per user, the current average rate, a dual variable, or the queue length, which is cheap to store and simple to update. They assume that the mobile services have a high tolerance for delay and that user positions and activity of users varies only slowly. Establishing long-term fairness by means of the methods described may lead to unacceptable periods without service for some users.
An extreme way to avoid this, is to establish fairness in each of the time-slots, for example for
or max-min fairness
As the current rate region is known, the maximization can be efficiently solved by suitable methods.
However, establishing a fair resource allocation in each time-slot independently may be too restrictive and lead to a loss in efficiency. Depending on the application, several consequent timeslots without service might be acceptable, but service needs to be provided within a fixed time window. A possible solution is predictive scheduling [16-22].
The idea is that, although they might be erroneous, estimates of future channel states might be beneficial. The resulting schedulers are no more memoryless and in general regard a certain horizon of past rate allocations (look-behind) and predictions of future channel states (look-ahead). For this time frame they maximize a utility or the expectation of the utility over several subsequent (potentially overlapping) time frames. So the gain of predictive scheduling comes at the price of having higher computational complexity.
For orthogonal access systems there is a direct connection between the data rate of the user and the channel state. This is no more true for advanced physical layer techniques, for example MU-MIMO, where by choosing the transmission strategies, for example transmission powers or beamformers, a trade-off between the user rates can be made. State of the art methods for predictive scheduling [16-22] are intended for orthogonal access systems and do not generalize to advanced multi-user systems.
Hence, for complex systems, the well-known methods are either too complex or too computationally expensive or do not result in the optimum solution with respect to a certain utility, such as a fair allocation utility.
According to an embodiment, an apparatus for scheduling transmission resources to users served by a base station equipped with a plurality of antennas may have: a predictor for predicting rate regions for one or more future time slots based on rate regions for one or more past time slots to acquire one or more predicted rate regions; and a processor for calculating the transmission resources for the users for a current time slot using scheduled transmission resources for the one or more past time slots, a rate region for the current time slot and the one or more predicted rate regions.
According to another embodiment, a method of scheduling transmission resources to users served by a base station equipped with a plurality of antennas may have the steps of: predicting rate regions for one or more future time slots based on rate regions for one or more past time slots to acquire one or more predicted rate regions; and calculating the transmission resources for the users for a current time slot using scheduled transmission resources for the one or more past time slots, a rate region for the current time slot and the one or more predicted rate regions.
According to another embodiment, a computer program may have a program code for performing, when running on a computer, the method of scheduling transmission resources to users served by a base station equipped with a plurality of antennas, which method may have the steps of: predicting rate regions for one or more future time slots based on rate regions for one or more past time slots to acquire one or more predicted rate regions; and calculating the transmission resources for the users for a current time slot using scheduled transmission resources for the one or more past time slots, a rate region for the current time slot and the one or more predicted rate regions.
The present invention is based on the finding that a predictor for predicting the rate regions for one or more future time slots based on rate regions for one or more past time slots is used to obtain one or more predicted rate regions. Then, the processor for calculating the transmission resources for the users for a current time slot uses the scheduled transmission resources for the past one or more time slots, a rate region determined for the current time slot and the one or more predicted rate regions output by the predictor.
Therefore, in contrast to estimating the channel gains for the next time slots, the present invention relies on estimating the rate regions rather than the channel gains for the future time slots. Particularly for the advanced multi-user systems, where advanced physical layer techniques are used such as multi-user MIMO, the channel gains cannot be directly translated into rate regions. Particularly for multi-user systems with adaptive modulation and coding, the data rates of the users are coupled and in theory infinitely many rate configurations can be provided. Hence, for such advanced multi-user systems, the complex interdependence of the user rates is a significant difference and therefore the present invention does not rely on the prediction of channel gains, but on the prediction of rate regions in order to shortcut the problem of translating channel gains for future time slots into rate regions for future time slots. Due to the fact that any prediction of channel gains are not required anymore in accordance with the present invention, one does not have to predict channels anymore. In a preferred embodiment, an opportunistic and/or fair resource allocation for multi-user systems is reduced to the formulation as an optimization problem. For proportional fairness, this system maximizes the sum of logarithmic average rates in an example. For the proportional fair sharing algorithm, the interpretation as utility maximization and proof for asymptotic optimality is existent. Hence, for scenarios where the direct connection between channel gains for next time slots and a prediction of the rate regions no longer exist, the present invention can nevertheless provide an improved asymptotic optimum solution for the scheduling of transmission resources. Particularly for advanced multi-user systems, by choosing the transmission strategies, for example transmission powers or beamformers, a trade-off between the user rates can be made. Statistical models of the channels might not be available and prediction methods require additional computational resources. Actually, in accordance with the present invention, one is not interested in the actual channel realizations in the future time slots, while in preferred embodiments a prediction of channels in the current time slot may be used to improve the quality of a delayed channel feedback. Instead, in accordance with the present invention, one is interested in the resulting user rates achievable in the future. Even if one could accurately predict the channels, it is not clear if one would invest the computational complexity that is useful for incorporating the channel prediction. The rate region prediction, however, does not require a channel prediction and the complex translation from predicted channels into predicted rate regions.
The present invention therefore relies on the prediction of the achievable user rates instead of predicting channels. For the prediction of rate regions, the information of the rate regions observed so far which can be given by the channel state information are used and it is assumed that these observations are representative of the future. Depending on the implementation, one prediction concept is based on the complete description of the previous rate regions which means the channel state information has to be stored and/or the past rate regions have to be stored. A further concept is based on inner approximations of previous rate regions, which drastically reduces memory requirements and allows for a specific implementation.
Further embodiments have a strong focus on cellular networks with base station cooperation, where a major concern is to keep the coordination overhead low. This means that an exchange of channel state information and/or channel state predictions should be avoided. For this scenario, particularly the concept based on the inner approximations may be specifically attractive, as the approximations are of much lower dimensionality, simple to exchange in a standardized way and lead to implementations with small coordination overhead.
For the predictive scheduling, the true average rate region is replaced by an approximation that depends on past allocations and the prediction. The number of past (look-behind) and future (look-ahead) time slots to be considered in the approximation can be configured and their influence can be adjusted by weights in preferred embodiments. The rate allocations of the previous time-slots cannot be changed and are assumed to be fixed. The current rate region is known, however the rate regions for the future time slots are not known and are therefore replaced by predictions. Under these assumptions, the rate allocation in the current time slot can be found by solving an optimization problem relying on past resource allocations, predicted rate regions and the current rate region for the current time slot.
In preferred embodiments, this optimization problem is implemented using the simplicial decomposition algorithm. This algorithm is modified to inner approximate all known rate regions that constitute the approximated rate region individually instead of inner approximating the estimated rate region. This has several advantages. As the problem has to be solved in every time-step, the inner approximations of the past time-slots can be reused.
These inner approximations are preferably directly used as representation of the old rate regions to form a prediction of future rate regions. Hence, it is avoided to store the much higher dimensional channel state information, although this would be possible as well. This is particularly important for multi-cell scenarios, where multiple base stations are coordinated and the exchange of channel state information should be avoided. Instead, it is preferred to consider distributed solutions with minor overhead that exchange the inner approximations (or some information derived from these), where the information exchange could be standardized to allow for interoperability.
Preferred embodiments of the present invention therefore provide an efficient concept for the prediction of rate regions, a predictive multi-user scheduler and/or an efficient implementation of the scheduler and the prediction.
The present invention provides a response to the ever increasing demand for higher data rates in wireless communications. Particularly, for opportunistic resource allocation or advanced physical layer techniques such as multi-user MIMO, the present invention is specifically advantageous. An efficient operation of a wireless network requires a balance between maximizing throughput and having a fair rate allocation among the users. Establishing long-term fairness by state of the art may lead to unacceptable periods without service, which can be avoided by predictive scheduling. Known methods for predictive scheduling are not applicable to advanced physical layer techniques with adaptive modulation and coding. Hence, the present invention relies on a predictive scheduler for advanced multi-user communication systems.
In other words, the prediction of the rate regions for future time slots allows that one can handle the requirement for fair allocation in an efficient way. For example, when a certain user is located quite far apart from the base station, this user typically does not have a very good transmission channel. Hence, this user will not get a high data rate or a high number of transmission resources under the scenario of optimizing the maximum throughput. However, the other requirement for fair allocation will increase the weight for this user more and more, i.e. over each time slot where the user again got a quite small number of transmission resources. By increasing the weights for this user, however, a situation will come where the scheduler actually schedules an increasing number of transmission resources to this user in order to fulfill the fairness requirement. However, this will result in a reduction of the overall throughput, since the allocation to the user with a not very good channel heavily affects the other users with good channels, which are situated more closely to the base station. Now, the present invention allows to “play” with the future of this channel. When there is a trend which is picked up by the prediction, which shows in the direction that the user is coming closer to the base station or the rate region where this user is located improves over time, a decision can be found saying that even though the user's weights actually force the resource allocator to now give this user a channel, to actually not do this for the current time slot or the future time slot, but probably for the next future time slot where the prediction indicates that the user is getting a better channel than before. Hence, the present invention allows that the transmission resource allocation actually waits one or more future time slots until the user is provided with more transmission resources in the hope of or with the help of the prediction that the user's channel will increase in the future and therefore the maximum throughput is higher as if the user would have been scheduled transmission resources even though the user had a bad channel. Hence, the prediction of rate regions allows to not only acknowledge the past and the current situation, but also the future situation and depending on the time extension of the prediction, i.e. how many future time slots are predicted, an optimum compromise between complexity, maximum overall throughput and fair allocation is obtained.
Embodiments of the present invention will be detailed subsequently referring to the appended drawings, in which:
Base stations are, for example, illustrated in
Furthermore,
The predictor 10 has an associated storage 26 for storing preferably parameterized rate regions from the past in a particularly efficient implementation. However, the storage 26 may store other representations of past allocations, but it is preferred to actually store either parameterized or non-parameterized rate regions from the past in order to generate from these rate regions from the past the rate regions for the one or more future time slots.
Preferably, a proportional fairness utility U is used as illustrated in
Furthermore, in order to assure fairness, the users are provided with weights as indicated at 36. As can be seen in the lower equation of item 36, the weights are inverse proportional to the user k′ rate. If the user received a higher rate in the past, the user will receive a low weight for the future and alternatively, if the user k received a low rate for the past, the user will receive a high weight. The weight influences the optimization problems solution so that users with high weights are preferred over users with low weight. This procedure makes sure that a certain user which has never received a high number of transmission resources in the past receives an even more increasing weight and at some in time the weight will he so high that the user is preferred over other users, although allocation of resources to this user violates the (unweighted) utility at 35, which is based on a maximization of the individual rates.
Predictive scheduling for orthogonal access systems, is based on estimating the channel gains for the next times-slots, which directly provides a prediction of the rates. This direct connection is not given for advanced multi-user systems.
For the prediction of the rate regions one uses information (the channel states) of the so far observed rate regions R[0], . . . , R[t] and assume that these observations are representative for the future. There are two specific methods to predict the rate region {tilde over (R)}[p] of a future time-slot p>t (the weighted sum of sets is defined as
prediction based on complete rate regions
prediction based on inner approximation
where RI[τ′] R[τ′] is an inner approximation formed by the convex hull of boundary points of R[τ′].
The approximation weights apτ′ can be used to adjust the influence of the past rate regions. A typical choice of the weights is to perform an average of a certain number of past rate regions but other choices, for example matched to the statistical properties and/or expected estimation errors, are possible.
For the predictive scheduling the true average rate region is replaced by an approximation. Based on the prediction we can define an approximate rate region
The variables Bt and Pt configure how many of the past (look-behind) and future (look-ahead) time-slots are considered in the approximation. The rate allocations r[t+1], . . . , r[t−Bt] of the Bt previous time-slots cannot be changed and are assumed to be fixed. The rate region R[t] is known. However, the rate regions of the Pt future time-slots are not known and are therefore replaced by predictions {tilde over (R)}[t+1], . . . , {tilde over (R)}[t+Pt]. The influence of each time-slot on the performance can be adjusted by the weights wτ, τ=t−Bt, . . . , t+Pt, and are usually chosen to match the definition of the average throughput considered. Under these assumptions the rate allocation in the current time-slot r[t] ∈ R[t] can be found by solving an optimization problem:
Contrary to problem (1.1) problem (1.7) can be solved casually. Given a solution
one obtains the rate configuration for the current time-slot: r[t]=r*[t]. Pit is noted that the rate allocations calculated for future time-slots are only virtual and will be recalculated in the next time-slot.
Subsequently, reference is made to
Problem (1.7) is typically solved by a sequence of weighted sum-rate optimizations, for ex-ample the simplicial decomposition algorithm [23, 24]. The simplicial decomposition algorithm consists of two steps, the column generation 44 and a master problem 46 for updating an estimate {tilde over (r)}i of the optimal solution. The master problem forms a convex combination of the previously generated columns and the new column to improve the estimate. The column generation in the i-th step is
i.e., one solves a weighted sum-rate (WSR) maximization. The master problem is
and the solution provides the update for the estimate {tilde over (r)}i. The master problem has simple constraints (the convex hull can be explicitly parameterized) and can be solved by standard methods for mathematical programming.
In the following, a more efficient solution method is provided, considering that similar problems are solved for several consequent time-slots. We use a generalized version of the simplicial decomposition method, see [25] for details. The algorithm uses an individual inner approximation of all rate regions involved, these can be any subset of all so far observed rate regions R[0], . . . , R[t]. Any combination of the inner approximations RI[0], . . . , RI[t] provides again a valid inner approximation, and we can therefore use the following master problem:
In the column generation step one enlarges the inner approximations. Due to the linearity of the objective, the WSR maximization in (1.8) can be decoupled into a subproblem per rate region involved. The solution is a weighted sum of the optimizers in each rate region. The enlargement of the inner approximation RI[t] of the current rate region R[t] is found by
In case the prediction based on complete rate regions is used, we also update the inner approximation of R[τ′]∀τ′=0, . . . , t−1 by
but only if R[τ′] is relevant for the prediction, that is Σp=t+1t+P
One has to consider that the master problem is very similar for several consequent time-slots, as the weights may change, but the inner approximations stay valid. This can be used to provide a more efficient implementation, by reusing previous inner approximations. In fact, this is also the intuition behind the prediction based on inner approximation. In this case one does not use the previous rate regions but the inner approximations obtained while running the simplicial decomposition algorithm. This means the column generation step operates only on the current rate region R[t]. Therefore, the complexity of the predictive algorithm is roughly the same as fairness per slot, given by (1.5) or (1.6), that also require multiple WSR maximizations and a master problem. It is however higher than for the memoryless algorithms that solve a single WSR maximization and have a closed form update for the variables they track.
Considering a multi-cell network, exchanging inner approximations of rate regions might be used to obtain a distributed solution for base station cooperation with minor overhead.
Subsequently, reference is made again to
Subsequently, reference is made to
Now, as illustrated in
Subsequently, reference is made to simulation results, particularly with reference to
Although the method is applicable to all three definitions of the average throughput, see
The channel coefficients of a user k depend on long term shadow fading that is log-normal distributed with mean μk and variance σk. The shadowing is independent for the users but the same for all channel coefficients of a user. The shadow fading is assumed to be constant within the window size regarded. The time variance of the channel coefficients is due to independent microscopic fading (Rayleigh fading), that depends on the mobility of the users.
Two scenarios are shown: one with homogeneous users where σk=0 dB for all users and one with heterogeneous users, where σk=4 dB for all users. The shadow fading mean is 0 dB in all cases.
An average over 100 windows is used, where channel realizations within one window are correlated, the realizations in different windows are independent. All users have an infinite backlog of traffic but the window length T is used to guarantee a fair share of service within a finite time window.
As a reference are used:
Maximum Throughput
Maximum throughput can be easily achieved causally by simple maximizing the sum-rate in each time-slot.
Max-Min Fairness
For max-min fairness we maximized the minimum rate of all users in every time-slot, which can be done by solving a utility maximization problem.
Non-Causal Upper Bound
The non-causal upper bound is found by assuming all channel realizations are known in advanced. In this case the utility optimal schedule, i.e., the optimal rate allocation for each time-slot can be computed.
Proportional Fair Per Slot
In this case proportional fairness is established in every time-slot, ignoring the past allocations and without considering a prediction.
Gradient Scheduler
Our numerical simulations show that the gradient scheduler clearly outperforms the stochastic subgradient and the queuing method and we therefore do not include them.
Predictive Proportional Fair Multi-User Scheduler (P-PF)
To evaluate performance the definition of the T normalized Doppler frequency is used as introduced in [20]. A small normalized Doppler frequency means that the channels hardly varies within the application time window, leaving little gains for the scheduler from being opportunistic.
The results for the heterogeneous scenario are shown in
Subsequently, further aspects of the present invention are discussed with reference to
The present invention provides a concept for multi-cell MIMO communications incorporating robust cooperative transmission strategies. Particularly, the robust cooperative transmission allows that users may be disconnected for several consequent transmission intervals, the goal is a robust strategy with short term fairness and the solution is a predictive multi-user scheduling.
The predictive multi-user scheduling relates to a downlink transmission for stochastic networks with time variant channels and for advanced multi-user transmission (MU-MIMO). Gains can be obtained by opportunistic resource allocation, but one can also observe a starvation of users. Hence, there has to be a trade-off between maximum throughput on the one hand and fairness on the other hand. The goal is an opportunistic and fair resource allocation. Proportional fairness in multi-user MIMO systems or advanced multi-user systems is related to the connection of proportional fairness and utility maximization. In an embodiment, the gradient algorithm is preferred serving as a generalized proportional fairness scheduling for multi-user systems.
Furthermore, the future channels H[4], H[5] are unknown and the channels H[3] and the current time slot are known. Hence, only the rates r[3], which are in rate region for the current slot 3, can be allocated.
However, this approach is problematic due to the fact that an MIMO channel prediction has to be performed and a high complexity is required for performing the complex mapping of the channel state to the rate region as indicated at the right of
In order to address all of these problems, the inventive concept as illustrated in
However, other prediction methods can be applied as well. For example, there can be a prediction as known from linear prediction coding of speech, where prediction coefficients are calculated and the predicted rate region is a weighted sum of the earlier rate regions, where the weights for the sum are determined by a linear prediction concept. Additionally, other ways of calculating the rate regions for the future based on the rate region for the past can be used as well.
Hence, the optimization problem solved by the present invention is that the utility U(r) is maximized where the sum of the rates is variable and the rates for the current and the two future time slots are variable as well. The constraint is that the transmission resources have to be in the rate regions for the current and the future time slots as indicated in
Furthermore, it is preferred to actually calculate the user's transmission resources r1 and r2 using the logic discussed in the context of
Furthermore, it is preferred to store the parameterized approximation of the rate region by the points indicated by the transmission resources r1, r2 and the axis points 37a, 37b. In implementations, however, the points 37a, 37b are not necessarily required, since all of the information is provided in the complete transmission resources found by an optimization in a certain rate region. Hence, the transmission resources themselves form an approximation of the rate region and it is preferred to use the earlier assigned transmission resources as the approximations for the rate regions to be used by the predictive scheduler on the one hand and to be used by the predictor on the other hand in a particularly efficient implementation.
Hence, the rate regions are approximated by the optimization results of earlier steps which are located on boundary points of the rate regions. It has been found that using the inner approximations instead of the rate regions themselves provides a great advantage with respect to complexity, but is not decisive with respect to the accuracy of the determination of the transmission resources for the current time slot.
The present invention therefore provides predictive scheduling in the rate space. Preferably, inner approximations of rate regions are stored, and the prediction is used based on inner approximations and therefore the complexity and storage requirements can be drastically reduced and only a marginal increase in complexity compared to a gradient scheduler is obtained, but a significant gain with respect to throughput and fairness performance.
Regarding
Although some aspects have been described in the context of an apparatus, it is clear that these aspects also represent a description of the corresponding method, where a block or device corresponds to a method step or a feature of a method step. Analogously, aspects described in the context of a method step also represent a description of a corresponding block or item or feature of a corresponding apparatus.
Depending on certain implementation requirements, embodiments of the invention can be implemented in hardware or in software. The implementation can be performed using a digital storage medium, for example a floppy disk, a DVD, a CD, a ROM, a PROM, an EPROM, an EEPROM or a FLASH memory, having electronically readable control signals stored thereon, which cooperate (or are capable of cooperating) with a programmable computer system such that the respective method is performed.
Some embodiments according to the invention comprise a non-transitory data carrier having electronically readable control signals, which are capable of cooperating with a programmable computer system, such that one of the methods described herein is performed.
Generally, embodiments of the present invention can be implemented as a computer program product with a program code, the program code being operative for performing one of the methods when the computer program product runs on a computer. The program code may for example be stored on a machine readable carrier.
Other embodiments comprise the computer program for performing one of the methods described herein, stored on a machine readable carrier.
In other words, an embodiment of the inventive method is, therefore, a computer program having a program code for performing one of the methods described herein, when the computer program runs on a computer.
A further embodiment of the inventive methods is, therefore, a data carrier (or a digital storage medium, or a computer-readable medium) comprising, recorded thereon, the computer program for performing one of the methods described herein.
A further embodiment of the inventive method is, therefore, a data stream or a sequence of signals representing the computer program for performing one of the methods described herein. The data stream or the sequence of signals may for example be configured to be transferred via a data communication connection, for example via the Internet.
A further embodiment comprises a processing means, for example a computer, or a programmable logic device, configured to or adapted to perform one of the methods described herein.
A further embodiment comprises a computer having installed thereon the computer program for performing one of the methods described herein.
In some embodiments, a programmable logic device (for example a field programmable gate array) may be used to perform some or all of the functionalities of the methods described herein. In some embodiments, a field programmable gate array may cooperate with a microprocessor in order to perform one of the methods described herein. Generally, the methods are preferably performed by any hardware apparatus.
While this invention has been described in terms of several embodiments, there are alterations, permutations, and equivalents which fall within the scope of this invention. It should also be noted that there are many alternative ways of implementing the methods and compositions of the present invention. It is therefore intended that the following appended claims be interpreted as including all such alterations, permutations and equivalents as fall within the true spirit and scope of the present invention.
Number | Date | Country | Kind |
---|---|---|---|
12156837.2 | Feb 2012 | EP | regional |