JP4173122B2 - Communication control device, communication control method, node, and communication system - Google Patents

Description

  The present invention relates to a communication control device, a communication control method, a node, and a communication system. For example, the present invention is distributed in a space such as a system composed of a plurality of devices connected to a sensor network or a LAN (Local Area Network). In addition, when a large number of nodes or nodes installed in a mobile body perform data communication with each other, communication data collision due to radio wave interference or the like is avoided.

  As a system for enabling data communication without a collision between a plurality of nodes distributed in space, there are a TDMA system, a CSMA (CSMA / CA and CSMA / CD) system, and the like (see Non-Patent Document 1). .

  In the CSMA method, a node to be transmitted confirms whether or not another node is communicating based on the presence of a carrier (frequency), and transmits when communication is not being performed. However, in the case of the CSMA method, as the number of nodes that generate traffic increases, the overhead increases, and a reduction in communication efficiency is inevitable.

In the TDMA system, different time slots are assigned to each node, and each node performs data transmission in the time slot assigned to itself. In the TDMA system, there are cases where a node used for communication changes dynamically. A node (management node) dynamically assigns a time slot to each node.
Matsushita Atsushi, Nakagawa Masao, “Wireless LAN Architecture”, Kyoritsu Shuppan, 1996, p. 47, 53-59, 69

  However, in the case of the TDMA system, if the management node that performs time slot allocation fails, the entire communication system goes down. Also, the process of dynamically reassigning time slots to each node is complicated, and it may not be possible to respond quickly to changes in the situation. Furthermore, in the case of the TDMA system, the width of the time slot itself cannot be changed.

  Therefore, a highly flexible communication control device, communication control method, node, and communication system are desired that enable each node to execute effective communication without the management node instructing each node of communication timing.

1st this invention was equipped with the communication timing calculation means which determines the timing of the data transmission from a self-node based on the state of the phase changed internally mounted in each of the some node which comprises a communication system. In the communication control device, the communication timing calculation means is configured according to a predetermined time development rule based on reception of a state variable signal from a neighboring node reflecting a phase representing a data transmission timing of the neighboring node. The phase calculation unit that changes the phase state of the node, the phase difference between the own node and the neighboring node is observed, and based on the phase difference, the data transmission timing from the own node and the data transmission timing from the neighboring node Collision rate calculation unit that calculates the collision rate, and stress values according to the collision rate are accumulated over time, and according to the accumulated stress values, Together and a serial phase calculation section stress response function value generating unit which occur a phase shift of random magnitude and are time evolution rule using continuous phase as a state variable signal that reflects the phase of the node The phase calculation unit includes a phase signal communication unit that continuously transmits a signal, receives a phase signal transmitted by a neighboring node, and enables the phase calculation unit to use the phase of the neighboring node .

  The node according to the second aspect of the present invention includes the communication control apparatus according to the first aspect of the present invention.

  A communication system according to a third aspect of the present invention is characterized in that a plurality of nodes according to the second aspect of the present invention are distributed and arranged.

The fourth aspect of the present invention is a communication control including a communication timing calculation step for determining a timing of data transmission from the own node based on a phase state to be changed internally, which is executed by each of a plurality of nodes constituting the communication system. In the communication timing calculation step, the local node receives the state variable signal from the neighboring node reflecting the phase indicating the data transmission timing of the neighboring node, and receives the state variable signal from the neighboring node according to a predetermined time development rule. The phase calculation sub-process for changing the phase state and the phase difference between the own node and the neighboring node are observed, and based on the phase difference, the timing of data transmission from the own node and the data transmission timing from the neighboring node Collision rate calculation sub-process for calculating the collision rate and the stress value corresponding to the collision rate are accumulated over time, and the accumulated stress value Flip and, together with and a said phase calculating substep stress response function value generating sub step of occur a phase shift of random magnitude and are time evolution rule used, as a state variable signal that reflects the phase of the node Continuously transmitting a continuous phase signal and receiving a phase signal transmitted by a neighboring node, and the phase calculation sub-process includes a phase signal communication process for making the phase of the neighboring node available. Features.

  According to the communication control device, the communication control method, the node, and the communication system of the present invention, each node interacts with neighboring nodes through the exchange of state variable signals even when there is no centralized management node. Communication can be executed by allocating time slots in an autonomous and distributed manner, and a time slot having a minimum size required for data transmission cannot be obtained, and data transmission cannot be performed substantially. The occurrence of a node that falls into a state can be greatly suppressed, and the stability and communication efficiency of data communication between nodes can be improved.

(A) First Embodiment Hereinafter, a first embodiment of a communication control device, a communication control method, a node, and a communication system according to the present invention will be described in detail with reference to the drawings. Note that this first embodiment assumes a system in which a large number of nodes distributed in space exchange data with each other wirelessly, such as a sensor network or an ad hoc network.

  In the first embodiment, each node generates an impulse signal, and by effectively detecting an impulse signal generated by a node other than itself, the nodes interact with neighboring nodes, and time is distributed autonomously. The assignment of slots is determined.

  The communication system according to the first embodiment includes a plurality of nodes distributed and arranged in a space that exchange data via a wireless line. Here, the first embodiment assumes that the position of each node hardly changes during communication.

(A-1) Nodes of First Embodiment Each node has a detailed configuration shown in the functional block diagram of FIG. In FIG. 2, the node 10 includes an impulse signal reception unit 11, a communication timing calculation unit 12, an impulse signal transmission unit 13, a tuning determination unit 14, a data communication unit 15, and a sensor 16. As a communication control device, impulse signal receiving means 11, communication timing calculating means 12, impulse signal transmitting means 13 and tuning determination means 14 are constituent elements.

  The impulse signal receiving means 11 receives as an input impulse signal Sin11 an output impulse signal (not including destination information) transmitted by a nearby node (for example, another node existing in a range where a transmission radio wave of the node reaches). Is. Here, the impulse signal is transmitted and received as a timing signal, and has an impulse shape such as a Gaussian distribution shape (in addition, the impulse signal may hold some data information). The reception impulse signal Spr11 may be a waveform-shaped signal of the input impulse signal Sin11 or a signal regenerated.

  Although details will be described later, the communication timing calculation means 12 forms and outputs a phase signal Spr12 that defines the communication timing at the node based on the received impulse signal Spr11 (note that the received impulse signal Spr11). Even if there is no signal, the phase signal Spr12 is formed and output). Here, assuming that the phase value at time t of the phase signal Spr12 of the node i is θi (t), the communication timing calculation means 12 is based on the received impulse signal Spr11, as described later, the phase signal Spr12 (= θi (t)) is changed with a nonlinear vibration rhythm. This change in phase signal is a result of realizing a non-linear characteristic in which neighboring nodes are opposite in phase (vibration phase is inverted) or other phase, and trying to avoid collision using that characteristic. It is. That is, an appropriate time relationship (time difference) is formed so that the transmission timing of the output impulse signal Sout11 between neighboring nodes does not collide.

  The meaning of the function of the communication timing calculation means 12 will be described in detail with reference to FIGS. 3 and 4. The state changes shown in FIGS. 3 and 4 are also related to the function of the impulse signal transmission means 13.

  3 and 4 show a relationship formed between a target node (own node) and a neighboring node (another node) when attention is paid to a certain node, that is, a phase relationship between respective nonlinear vibration rhythms. Shows how the changes over time.

  FIG. 3 shows a case where there is one neighboring node j for the node of interest i. In FIG. 3, the motion of the two mass points rotating on the circle represents the nonlinear vibration rhythm corresponding to the node of interest and the neighboring nodes, and the angle of the mass point on the circle represents the value of the phase signal at that time. Yes. The motion of the point where the rotational motion of the mass point is projected on the vertical or horizontal axis corresponds to the nonlinear vibration rhythm. By the operation based on the expression (1) described later, the two mass points try to be in opposite phases to each other, and even if the phases of the two mass points are close to each other in the initial state as shown in FIG. At the same time, the state (transient state) shown in FIG. 3B is changed to a steady state in which the phase difference between the two mass points is approximately π as shown in FIG. 3C.

  Each of the two mass points rotates with the natural angular frequency parameter ω as a basic angular velocity (corresponding to a basic velocity that changes its own operation state). Here, when an interaction based on transmission / reception of an impulse signal occurs between nodes, these mass points change (slow / slow) angular velocities, respectively, and eventually reach a steady state in which an appropriate phase relationship is maintained. This operation can be regarded as forming a stable phase relationship by repelling each other while the two mass points rotate. In the steady state, as will be described later, when the output impulse signal Sout11 is transmitted when each node is in a predetermined phase (for example, 0), the transmission timings at the nodes form an appropriate time relationship. It will be.

  FIG. 4 shows a case where there are two neighboring nodes j1 and j2 for the node of interest i. Even in the case where there are two neighboring nodes, a stable phase relationship (stability related to temporal relationship) is formed by repelling each other while rotating the respective mass points as described above. The same applies to the case where the number of neighboring nodes is three or more.

  The formation of the above-described stable phase relationship (steady state) has a very adaptive (flexible) property with respect to changes in the number of neighboring nodes. For example, it is assumed that one neighboring node is added when there is one neighboring node with respect to the node of interest and a stable phase relationship (steady state) is formed. The steady state once collapses, but after passing through the transient state, a new steady state in the case where there are two neighboring nodes is reformed. In addition, when a neighboring node is deleted or does not function due to a failure or the like, an adaptive operation is performed in the same manner.

  The communication timing calculation unit 12 outputs the obtained phase signal Spr12 (= θi (t)) to the impulse signal transmission unit 13, the tuning determination unit 14, and the data communication unit 15.

  The impulse signal transmission unit 13 transmits the output impulse signal Sout11 based on the phase signal Spr12. That is, when the phase signal Spr12 reaches a predetermined phase α (0 ≦ α <2π), the output impulse signal Sout11 is transmitted. Here, it is preferable that the predetermined phase α is previously unified in the entire system. In the following description, it is assumed that α = 0 is unified throughout the system. In the example of FIG. 3, since the mutual phase signal Spr12 is shifted by π in a steady state between the node i and the node j, even if the entire system is unified to α = 0, The transmission timing of the output impulse signal Sout11 and the transmission timing of the output impulse signal Sout11 from the node j are shifted by π.

  The tuning determination means 14 is a “transient state” in which the transmission timing of the output impulse signal Sout11 performed between the own node and one or a plurality of neighboring nodes is “transient” (see FIGS. 3B and 4B). Alternatively, it is determined which state is the “steady state” (see FIG. 3C or FIG. 4C). The tuning determination unit 14 observes the generation timing of the reception impulse signal Spr11 (corresponding to the output impulse signal Sout11 of the other node) and the output impulse signal Sout11, and the time difference between the generation timings of a plurality of nodes that exchange the impulse signals. When it is stable over time, it is determined that the state is “steady state”. In the case of this embodiment, the tuning determination means 14 receives the phase signal Spr12 instead of the output impulse signal Sout11 as a signal for capturing the generation timing of the output impulse signal Sout11 from the own node. Yes.

  The tuning determination unit 14 performs the tuning determination by executing the following processes (a) to (d), for example.

  (A) The value β of the phase signal Spr12 at the generation timing of the reception impulse signal Spr11 is observed over one period of the phase signal Spr12. Here, as a result of the above observation, the obtained values β of the phase signal Spr12 are β1, β2,..., ΒN (0 <β1 <β2 <... ΒN <2π).

  (B) Based on the observed value β of the phase signal Spr12, the difference (phase difference) between adjacent values Δ1 = β1, Δ2 = β2-β1,..., ΔN = βN−β (N−1) Is calculated.

  (C) The above processes (a) and (b) are performed in units of the period of the phase signal Spr12, and the amount of change (difference) in the phase difference Δ in successive periods γ1 = Δ1 (τ + 1) −Δ1 ( τ), γ2 = Δ2 (τ + 1) −Δ2 (τ),..., γN = ΔN (τ + 1) −ΔN (τ) are calculated. Here, τ indicates a certain cycle of the phase signal Spr12, and τ + 1 indicates the next cycle of the phase signal Spr12.

  (D) When the above-mentioned change amount γ is smaller than the minute parameter (threshold value) ε, that is, when γ1 <ε, γ2 <ε,. To do.

  Note that a steady state may be determined when the conditions of γ1 <ε, γ2 <ε,..., ΓN <ε are satisfied over M cycles. As the value of M is increased, it is possible to determine “steady state” in a more stable state. Further, the “steady state” may be determined based on a part of the received impulse signal Spr11.

  The tuning determination means 14 is a data communication means in which the tuning determination signal Spr13 indicating the determination result and the minimum value β1 of the value β of the phase signal Spr12 at the generation timing of the reception impulse signal Spr11 are set as the slot signal Spr14 for each cycle of the phase signal Spr12. 15 is output. Note that the output of the minimum value β1 as the slot signal Spr14 is related to α = 0 as described above, and β applied to the slot signal Spr14 depends on the selection of the value of α. The value of varies.

  The node 10 has a function of relaying and transmitting data received from other nodes, and a data transmission function using itself as a transmission source.

  The sensor 16 is written as an example of the latter case. For example, the sensor 16 detects physical or chemical environment information Sin13 such as sound and vibration intensity, chemical substance concentration, temperature, etc., and obtains observation data Spr15 as data. This is output to the communication means 15.

  In the former case, the data signal (output data signal Sout12) transmitted by the neighboring node is received by the data communication means 15 as the input data signal Sin12.

  The data communication unit 15 transmits the observation data Spr15 and / or the input data signal Sin12 (including both cases) as an output data signal Sout12 to another node. When the tuning determination signal Spr13 indicates “steady state”, the data communication means 15 uses the term “time slot”, which is a time slot (not a fixed time interval assigned by the system, etc.) described later. When the tuning determination signal Spr13 indicates “transient state”, the transmission operation is stopped. Note that the output data signal Sout12 may have a transmission frequency in the same frequency band as the output impulse signal Sout11.

  The time slot is a period in which the phase θi (t) of the phase signal Spr12 is δ1 ≦ θi (t) ≦ β1-δ2. The start point of the time slot (the value of the phase signal at that time is δ1) is the timing at which the transmission of the output impulse signal Sout11 is completed, and the end point of the time slot (the value of the phase signal at that time is β1-δ2). ) Is a timing that is slightly offset δ2 before the timing of the first reception impulse signal Spr11 for each period of the phase signal Spr12. δ1 and δ2 are an impulse signal (including both the case where the transmission source is the own node and the case of the other node) and the data signal (when the transmission source is the own node, the other node) in the wireless space near the node 10 Phase width corresponding to a very short time to compensate for the absence of both at the same time. For example, δ1 and δ2 are experimentally determined under the installation state of the node 10.

  For example, in the “steady state” as shown in FIG. 3C, the node i starts transmitting the output impulse signal Sout11 from the phase θi of 0, and before the phase θi becomes δ1, the node i of the output impulse signal Sout11 The transmission is terminated, the phase θi starts to transmit the output data signal Sout12 from δ1, and when the phase θi becomes β1−δ2 (where β1≈π), the transmission of the output data signal Sout12 is terminated. Until θi becomes 0 again, transmission of the output impulse signal Sout11 and transmission of the output data signal Sout12 are stopped. The other node j also performs the same operation based on the phase θj. However, since the phase θi and the phase θj are substantially shifted by π, the transmission operation does not compete. The same operation is performed when the number of nodes is 3 or more, and transmission operations do not compete.

(A-2) Details of Communication Timing Calculation Unit 12 FIG. 1 is a block diagram showing a detailed configuration of the communication timing calculation unit 12. In FIG. 1, the communication timing calculation unit 12 includes a phase calculation unit 21, a collision rate calculation unit 22, an accumulated stress calculation unit 23, and a stress response function value calculation unit 24.

As described above, the communication timing calculation unit 12 performs calculation for determining the timing for transmitting the output impulse signal Sout11. The communication timing calculation means 12 performs calculation for determining the transmission timing using, for example, a mathematical expression that models nonlinear vibration such as Expression (1).

  Equation (1) is an equation that represents a rule (time evolution rule) that temporally changes the rhythm of nonlinear vibration of the node (node i) according to the input of the received impulse signal Spr11. Here, the received impulse signal Spr11 corresponds to the output impulse signal Sout11 transmitted by the neighboring node j.

  In the equation (1), the variable t represents time, and the function θi (t) represents the phase with respect to the nonlinear vibration of the node at time t. It is assumed that the function θi (t) always takes a value of section 0 ≦ θi (t) <2π by performing an operation of mod 2π (residue divided by 2π).

  Δθij (t) is a phase difference obtained by subtracting the phase θi (t) of the own node i from the phase θj (t) of the neighboring node j as shown in the equation (2). Assuming that the output impulse signal Sout11 is transmitted when each node has a phase θi (t) = 0, the phase difference Δθij (t) observable by the node i is the timing at which the output impulse signal is received from the neighboring node j. (At this time, θj (t) = 0) only, and once per cycle. At this time, the phase difference Δθij (t) is −θi (t) as shown in the equation (2). However, the phase difference Δθij (t) is assumed to take a value of section 0 ≦ Δθij (t) <2π for the sake of convenience by performing an operation of mod 2π on a value obtained by adding 2π. In the first embodiment, the phase difference Δθij (t) is an amount given as described above. In the above description, it is assumed that each node transmits the output impulse signal Sout11 when the phase θi (t) = 0. However, even if such an assumption is made, generality is not lost. Even if it is assumed that the output impulse signal Sout11 is transmitted at a phase other than 0, the same operation can be performed.

  ωi is a natural angular frequency parameter and represents a basic rhythm provided in each node. Here, as an example, it is assumed that the value of ωi is previously unified to the same value in the entire system.

  The function Pj (t) represents the received impulse signal Spr11 received from the neighboring node j. This function represents a timing signal (a signal that does not have data and simply transmits timing) having an impulse-like function shape such as a rectangle or a Gaussian distribution. The pulse width and amplitude value of the function Pj (t) are determined experimentally, for example.

  The function R (Δθij (t)) is a phase response function that expresses a response characteristic that changes the basic rhythm of the own node according to the input of the received impulse signal Spr11. N represents the total number of neighboring nodes existing in the spatial distance range in which the own node can receive the received impulse signal Spr11.

  The function ξ (Si (t)) accumulates stress when the relative phase difference between the self node and the neighboring node is small, and a phase shift (phase) with a random magnitude according to the accumulated stress value Si (t). It is a term that works to execute (state change). Here, the relative phase difference is an amount defined as follows. When the phase difference is Δθij and the relative phase difference is E, the value satisfies the expression (3).

When Δθij ≦ π, E = Δθij
When Δθij> π, E = 2π−Δθij (3)
Therefore, the function ξ (Si (t)) is a function expressing the response characteristic with respect to the accumulated stress value Si (t). Hereinafter, this function ξ (Si (t)) is referred to as a stress response function.

  In the first embodiment, even if there is no centralized management node, each node interacts with a neighboring node, and by changing the phase signal according to the model expressing the nonlinear vibration, the time slot is autonomously distributed. In order to avoid the occurrence of a node that stabilizes as a steady state even if the relative phase difference between its own node and neighboring nodes is very small It is characterized in that a stress response function ξ (Si (t)) is introduced into a model expressing vibration.

  Next, before explaining specific examples of the stress response function ξ (Si (t)) and the phase response function R (Δθij (t)) and their functions, the terms related to these functions are: (A) “Data transmission “Collision”, (A) “Collision rate (when the collision time is used as a reference)”, and (C) “Collision rate (when the number of collisions is used as a reference)” will be described.

(A) "Data transmission conflict"
(A-1) Let φc be the phase width corresponding to the minimum time slot size Wmin necessary for each node to transmit data. The phase width φc can be calculated as the product of Wmin and the natural frequency parameter ωi (φc = Wmin · ωi). Wmin is a constant parameter determined according to the application.

  (Ar 2) When the relative phase difference between each node and a neighboring node existing in the spatial distance range in which the impulse signal can be received is smaller than the phase width φc, “a data transmission collision has occurred”. Shall. If even one of the plurality of received impulse signals Spr11 has a relative phase difference smaller than the phase width φc, it is assumed that a collision has occurred.

(A) An example of a method for calculating the “collision rate (when the collision time is used as a reference)” As a function indicating whether or not a data transmission collision occurs at time t in the own node (node i), the function xi (T) is defined as shown in equation (4).

  The function xi (t) is a function that takes a value of 1 when a collision occurs at time t (collision) and 0 when it does not (else). Then, the cumulative collision time yi (t) between n periods is obtained by accumulating (time integrating) the value of the function xi (t) over n periods as shown in the equation (5). In the equation (5), Ti represents the period of the node i. The cumulative collision time yi (t) represents the sum of the times when the function xi (t) takes a value of 1 during n periods, and can be calculated by observing the value of the function xi (t).

  A value ci (t) obtained by normalizing the cumulative collision time yi (t) with the maximum cumulative collision time represents a temporal ratio of collision occurring in n cycles, and this is defined as a collision rate. Here, the maximum cumulative collision time is the maximum value of the cumulative collision time yi (t). Assuming that each node performs transmission using a time slot of size Wmin (= φc / ωi), the maximum value of the accumulated collision time during n periods is n · Wmin (= n · φc / ωi). Become. Therefore, the collision rate ci (t) can be calculated using the equation (6).

  However, since the phase θ of each node changes according to the equation (1), the period Ti can take a different value for each period. For this reason, the case where the cumulative collision time yi (t) exceeds the maximum cumulative collision time n · Wmin, that is, the case where the collision rate ci (t) exceeds 1, may occur. Here, when the collision rate ci (t) exceeds 1, it is treated as 1.

  In the above definition of the collision rate, as an example, a form that does not consider whether or not multiple collisions occur at the same time is shown. However, the method of defining the collision rate is not limited to the above method. For example, it is possible to use a method that considers the case where a plurality of collisions occur simultaneously. In the above definition of the collision rate, the collision rate is calculated based on the collision time, but it is also possible to calculate the collision rate based on the number of collisions.

(C) An example of a calculation method of “collision rate (when the number of collisions is used as a reference)” (S1) It is observed whether or not a collision has occurred in one cycle unit. However, even when a collision occurs a plurality of times within one cycle, only the presence or absence of the collision is considered as a problem, and the count is 1.

  (S2) Count the number of collisions that occurred during n cycles, that is, the cumulative number of collisions γ.

  (S3) A value obtained by normalizing the cumulative collision number γ with the maximum cumulative collision number (maximum number of collisions n that can occur during n cycles) is defined as a collision rate. That is, the collision rate ci (t) is defined by the following equation (7), and the collision rate is calculated using the equation (7).

ci (t) = γ / n (7)
Even when the collision rate is calculated using the definition formula of the equation (7), the collision rate may exceed 1 as in the case of the equation (6). When the collision rate exceeds 1, it is handled as 1. In the above definition of the collision rate, as an example, a form in which a plurality of collisions in one cycle is not taken into account (a form in which only the presence or absence of a collision is a problem) has been shown. Is possible.

  The definition of the collision rate described in the sections (a) and (c) above (when the collision time is used as a reference and when the number of collisions is used as a reference) is generalized for the collision time and the number of collisions and the amount of collision. Can be expressed as follows.

  A value obtained by observing the accumulated collision amount during n cycles and normalizing it by the maximum accumulated collision amount (maximum collision amount that can occur during n cycles) is defined as a collision rate.

  The collision rate calculation unit 22 in FIG. 1 calculates the collision rate ci (t) defined by the equation (6) or (7).

  Next, a specific example of the phase response function R (Δθij (t)) that is internally calculated by the phase calculation unit 21 and its function will be described.

The phase response function R (Δθij (t)) is defined by, for example, Expressions (8-1) to (8-3). FIG. 5 is an explanatory diagram showing the phase response function R (Δθij (t)) defined by the equations (8-1) to (8-3) in a graph.

  In the above equation, φd and α represent constant parameters, and their values are determined experimentally. The constant parameter φd takes a value greater than or equal to the minimum phase width φc necessary for data transmission (φd ≧ φc). The phase response function R (Δθij (t)) has a non-linear characteristic that changes the phase θi (t) of the own node i in the direction in which the repulsive force works with respect to the phase of the neighboring node j. When the relative phase difference Δθij (t) between the self node and the neighboring node is equal to or less than φd, the repulsive force is changed in a working direction. By providing the phase response function R (Δθij (t)) with such characteristics, the relative phase difference Δθij (t) between the self-node and the neighboring node is set to be larger than the phase width φc necessary for collision avoidance. Giving characteristics.

  However, the function form of the phase response function R (Δθij (t)) is of course not limited to the above. A function that gives a mechanical characteristic that makes the relative phase difference between the self-node and a neighboring node greater than or equal to the phase width φc can be realized by using various forms (function forms).

  Next, specific examples and functions of the stress response function ξ (Si (t)) obtained by the sequential calculation processing of the accumulated stress calculation unit 23 and the stress response function value calculation unit 24 will be described.

The stress response function ξ (Si (t)) is defined by, for example, the following expressions (9-1), (9-2), (10-1), (10-2), and (11). It is what is done.

  The function s (ci (t)) in the equation (11) is a function representing a stress value with respect to the collision rate ci (t) at time t. For example, it is realized as a function having such a characteristic that a larger stress value is shown as the collision rate ci (t) is higher. As an example, there is a function having a characteristic that shows a large stress value rapidly as the collision rate ci (t) increases by using a nonlinear function such as a sigmoid function.

  The function Si (t) defined by the equation (11) is a function indicating a value obtained by accumulating (time integration) the stress value s (ci (t)) at time t. The time interval for accumulation is from the time ts when the random phase shift is executed in accordance with the stress value Si (t) accumulated last time to the current time t. That is, the function Si (t) is reset once the random phase shift is performed, and repeats the operation of accumulating the stress value s (ci (t)) again from that time. The integration operation in the function Si (t) can be calculated as the sum of the stress values s (ci (t)) at each time when the time t is given in a discretized manner. The accumulated stress calculation unit 23 calculates accumulated stress Si (t).

  The function q (Si (t)) defined by the equation (10-1) or (10-2) is a function that returns a random value with a probability corresponding to the accumulated stress value Si (t). The value μ is returned with probability Si (t), and the value 0 is returned with probability 1-Si (t). The value μ is a random number within the interval ε ≦ μ <δ, and the values ε and δ are constant parameters determined experimentally.

  The stress response function ξ (Si (t)) is a function that evaluates the accumulated stress value Si (t) every n periods and returns a random value (μ or 0) with a probability corresponding thereto. The stress response function ξ (Si (t)) is calculated by the stress response function value calculation unit 24.

  Therefore, by introducing the stress response function ξ (Si (t)) to the nonlinear vibration model as shown in the above equation (1), (I) the accumulated stress value Si (t) becomes n. Evaluation is performed for each period, and (II) an operation in which a random phase shift is executed with a probability based on the evaluation value is realized. That is, as the stress value accumulation due to the collision is larger, the random phase shift is executed with higher probability. Except for a time that is an integral multiple of n cycles, the value of the stress response function ξ (Si (t)) is 0, and no random phase shift is performed. However, it should be noted that the stress accumulation is not limited to n cycles as described above, but from the time ts when the random phase shift was performed last time to the current time t. For example, even if the stress value s (ci (t)) is small, if the stress value s (ci (t)) is continuously accumulated over a long period of time of n cycles or more, the stress value accumulation Si (t) eventually becomes large and random. This means that a correct phase shift can be performed.

  The phase calculation unit 21 calculates the phase θi (t) determined by the equation (1) while applying the stress response function value ξ (Si (t)).

  Note that the communication timing calculation unit 12 that appropriately executes the above-described calculation may be realized by software, may be realized by hardware based on an electronic circuit group that executes the calculation, and It may be realized by mixing hardware.

  For example, the execution means of the operation shown in the equation (1) is Runge Kutta disclosed in Reference 2, “Toto Hayato,“ Science and Technology Computation Handbook on UNIX Workstations—Basic C Language Version ”, published by Science Inc.”. It can be implemented on a node as general software such as law. The Runge-Kutta method is one of the methods for calculating the change (time evolution) of a state variable using a difference equation (recursion formula) obtained by differentiating a differential equation (discretizing a continuous time variable t). is there.

(A-3) Effect of First Embodiment According to the first embodiment, each node interacts with a neighboring node through exchanging impulse signals even when there is no centralized management node. It is possible to execute communication by determining time slot assignment in an autonomous and distributed manner.

  Here, according to the first embodiment, it is not possible to acquire a time slot having a minimum size necessary for data transmission, and greatly suppress the occurrence of a node that falls into a state in which data transmission cannot be performed. It is possible to improve the stability and communication efficiency of data communication between nodes.

(B) Second Embodiment Next, a second embodiment of the communication control device, the communication control method, the node, and the communication system according to the present invention will be described in detail with reference to the drawings.

(B-1) Node of the Second Embodiment In the first embodiment, the mode in which the nodes transmit and receive the output impulse signal Sout11 to each other has been described. At that time, the output impulse signal Sout11 has been used as a means for realizing an interaction between nodes (a function that affects the mutual phase state).

  In the present invention, the means for realizing the interaction between the nodes is not limited to that of the first embodiment. For example, it is also possible to configure a form of interaction by directly transmitting and receiving the phase θi (t) in the first embodiment. That is, it is possible to operate even in a form in which changes in the operation state such as the phase state in each node are transmitted and received with each other as a continuous signal.

  In the second embodiment, interaction is achieved by continuously transmitting and receiving phase signals between a plurality of nodes. FIG. 6 is a block diagram illustrating a configuration example of a node in the second embodiment that continuously transmits and receives phase signals.

  The node 30 in the second embodiment includes a phase signal reception unit 31, an interference detection unit 32, a data communication unit 33, a communication timing calculation unit 34, and a phase signal transmission unit 35, which are omitted in FIG. A tuning determination means (14) and a sensor (16) similar to those of the first embodiment are also provided. The data communication means 33 is also the same as the corresponding configuration (15) of the first embodiment.

  The phase signal receiving means 31 provided in place of the impulse signal receiving means 11 of the first embodiment receives the input phase signal Sin31 which is a continuous signal transmitted from the other node and received by the node 30. Information contained in the signal is extracted. The input phase signal Sin31 defines the phase signal θj (t) and the interference band number at the other node j, and includes the phase signal θj (t) with respect to the continuous signal having the frequency determined by the interference band number. (For example, PM signal). The phase signal receiving means 31 gives the phase signal θj (t) at the other node j in the input phase signal Sin31 to the communication timing calculating means 34, and the reception state detection result (for example, the interference band number and its reception timing). This is given to the interference detection means 32.

  Based on the detection result of the reception state from the phase signal receiving unit 31, the interference detection unit 32 determines the communication detection result as to whether or not the interference band number in the output phase signal Sout31 is duplicated (interference) at a plurality of nodes. In addition to being provided to the calculation means 34, the interference detection result and the interference band number detected by the interference are provided to the phase signal transmission means 35.

  The communication timing calculation unit 34 forms the phase signal θi (t) of its own node based on the phase signal θj (t) of the other node, etc., in substantially the same manner as in the first embodiment. The communication timing calculation means 34 will be described in detail later.

  A phase signal transmission unit 35 provided in place of the impulse signal transmission unit 13 of the first embodiment transmits an output phase signal Sout31 including the phase signal θi (t) of the own node.

  Here, in the second embodiment, it is assumed that the output phase signal Sout31 and the output data signal Sout32 from the own node 30 are transmitted using different frequency bands. Further, the frequency band for transmitting the output phase signal Sout31 is subdivided into Nb (Nb: integer) bands.

  When a system having a plurality of nodes according to the second embodiment starts operation, each node executes an initial operation consisting of the following Step 1 to Step 4.

  (Step 1) Any one of the Nb frequency bands is randomly selected, and the output phase signal Sout31 is transmitted using the selected band.

(Step 2) Whether a plurality of neighboring nodes transmit the output phase signal Sout31 using the same frequency band is detected based on the reception state of the input phase signal Sin31. If the input phase signal Sin31 can be normally received in each frequency band, it is determined that interference has not occurred, and conversely, if reception has failed, it is determined that interference has occurred. Further, it is determined that the frequency band where carrier sense cannot be performed is unused. When there is a frequency band in which interference occurs, a number indicating the interference band is added to the output phase signal Sout31 of the own node and transmitted.
(Step 3) In the case of the following (a) or (b), the frequency band used by the own node for transmission of the output phase signal Sout31 is shifted with probability P, and is left as it is with probability 1-P. Here, the probability P is a constant parameter determined experimentally. The above-described shift of the band to be used is executed by randomly selecting from the unused bands at that time.

(a) When the number of the frequency band used by the own node for transmitting the output phase signal Sout31 matches the number indicating the interference band added to the input phase signal Sin31 received from the neighboring node
(b) When the number of the frequency band used by the own node for transmitting the output phase signal Sout31 matches the number of the band where the input phase signal Sin31 cannot be normally received from the neighboring node (Step 4) When there is no frequency band in which the reception of the input phase signal Sin31 has failed and no number indicating the interference band is added to all the received input phase signals Sin31, each of the neighboring nodes is different. It is determined that the output phase signal Sout31 has been successfully transmitted and received using the frequency band. Hereinafter, this state is referred to as a stable interaction state.

  When a stable interaction state is formed by the initial operation, the operation of the communication timing calculation unit 34 as described below is started. However, even if a stable interaction state is once formed, when a new node is added, the state may collapse and a crosstalk band may be generated. In this case, the process for forming the stable interaction state of Step 2 to Step 4 is executed again. However, only the newly added node executes the operations of Steps 1 to 4 above.

(B-2) Communication Timing Calculation Unit of Second Embodiment Next, a detailed configuration and function of the communication timing calculation unit 34 in the second embodiment will be described.

  Similarly to the communication timing calculation unit 12 of the first embodiment, the communication timing calculation unit 34 of the second embodiment also includes a phase calculation unit 41, a collision rate calculation unit 42, an accumulated stress calculation unit 43, and a stress response function value. After the calculation unit 44 is provided and the above-described stable interaction state is formed, it functions in substantially the same manner as in the first embodiment. The phase calculation unit 41 also performs an initial operation for forming a stable interaction state.

The communication timing calculation unit 34 of the second embodiment performs calculation for determining the timing of transmitting the output phase signal Sout31 as described above after the stable interaction state is formed. The communication timing calculation means 34 performs calculation for determining the transmission timing using, for example, a mathematical expression modeling nonlinear vibration such as Expression (12).

  In the case of the first embodiment, the determination is made based on the received impulse signal of the other node, whereas in the case of the second embodiment, the determination is made based on the phase signal of the other node. In this case, equation (12) is applied instead of equation (1). In Equation (12), K is a constant parameter, and its value is determined experimentally. The other functions and parameters represent substantially the same as those in the first embodiment (for example, there is a difference in whether the phase of another node is directly given or obtained from reception of an impulse signal). However, the explanation is omitted.

  When the communication timing calculation unit 34 of the second embodiment forms the phase signal θi (t) of its own node based on the equation (12), the difference from the first embodiment will be described below. Street.

  In the case of the second embodiment, regarding “data transmission collision”, it is assumed that each node performs data communication using a time interval in which its phase θi (t) is 0 ≦ θi (t) <φc. Then, when the phase θi (t) of the own node is 0 ≦ θi (t) <φc (φc is a phase width corresponding to the minimum required time slot size), 0 ≦ If there is one that takes the phase θj (t) <φc, a data transmission collision has occurred. In the second embodiment, the term “data transmission collision has occurred” is used in the above meaning. However, as in the first embodiment, in each node, when the relative phase difference with the neighboring node existing in the spatial distance range in which the phase signal can be received is smaller than the above φc, “data transmission collision” May be defined.

  Except for the above points, the second embodiment is the same as the first embodiment. That is, the calculation method of the collision rate ci (t) by the collision rate calculation unit 42 is the same as the calculation method by the collision rate calculation unit 22 of the first embodiment. In addition, the calculation method of the phase response function R (Δθij (t)) by the phase calculation unit 41 is the same as the calculation method by the phase calculation unit 21 of the first embodiment. Furthermore, the calculation method of the accumulated stress value Si (t) by the accumulated stress calculation unit 43 and the calculation method of the stress response function value ξ (Si (t)) by the stress response function value calculation unit 44 are the same as those in the first embodiment. This is the same as the method for calculating the corresponding element.

(B-2) Effect of Second Embodiment According to the second embodiment, each node can communicate with neighboring nodes through the exchange of continuous phase signals even when there is no centralized management node. In operation, communication can be executed by determining time slot allocation in an autonomous and distributed manner.

  Here, according to the second embodiment, it is not possible to acquire a time slot having a minimum size required for data transmission, and greatly suppress the occurrence of a node that falls into a state where data transmission cannot be performed. It is possible to improve the stability and communication efficiency of data communication between nodes.

(C) Other Embodiments In the description of each of the above embodiments, the modified embodiment has been referred to, but further modified embodiments as exemplified below can be given.

(C-1) In the first embodiment, the differential equation (1) is used to describe the processing in the communication timing calculation unit 12. However, the description using a differential equation is only an example of the expression that defines the calculation method. Here, an example of a description method using a differential equation (recurrence formula) obtained by differentiating the differential equation (1) (discretizing the continuous time variable t) will be described.

  In equation (13), the variable u is a variable representing discrete time taking a positive integer value. The meanings of the other symbols are the same as in the first embodiment, but it should be noted that they are all functions of the discrete time variable u. The variable Δt represents a time step, and the continuous time variable t and the discrete time variable u have a relationship of t = u · Δt. Expression (13) indicates that the value of the phase θi (u + 1) at the next time u + 1 is calculated from the value of the phase θi (u) at the time u.

Further, the above description method is a description method in which the time evolution of the phase θi (t) is discretized in the time axis direction. In addition to discretization in the time axis direction, it is also possible to use a description method in which the state variable is discretized, that is, the value of the phase θi (t) itself is discretized (quantized). In this case, the possible values of the phase θi (t) are M (M: natural number) discrete values. This can be realized, for example, using equation (14). In equation (14), the function quan (*) represents a value obtained by dividing the variable * by the quantization width w and ignoring the numerical value after the decimal point. Here, the quantization width w is a value obtained by dividing the dynamic range of the variable * (the range of possible values) by M. In the equation (14), the phase θi (u) represents a quantized phase value at the discrete time u.

  Similarly, for the expression (12) shown in the second embodiment, it is possible to use the description method in the form of difference or discretization as described above. Any of these operations can be implemented on the node as software.

(C-2) In the first and second embodiments, Equations (1) and (12) are shown as mathematical equations modeling nonlinear vibrations, and various modifications are possible as other description methods. It was mentioned above.

  However, the description method of the mathematical expression that models the nonlinear vibration that realizes the present invention is not limited to the above-described embodiments and the modified embodiments already mentioned. For example, it is also possible to use general models of non-linear vibrations and chaotic vibrations such as the Van der Pol equation disclosed in Reference 3 “Morikazu Toda, Shinsuke Watanabe,“ Nonlinear Mechanics ”, published by Kyoritsu Shuppan”. is there. Of course, the interaction between nodes can be realized both discretely (pulse-like) with respect to time and continuous. The Van der Pol equation is an equation that models a nonlinear vibration phenomenon that occurs in an electronic circuit. The operation based on the Van der Pol equation can be implemented on a node as hardware using an electronic circuit. Further, it is possible to implement on a node as software by using a general numerical calculation method such as Runge-Kutta method.

  The present invention does not depend on differences in the description method of individual models, such as discrete models related to time, operating state, interaction, etc., continuous models, and other models expressing specific vibration phenomena. It can be realized by using various models in which the operation state transitions according to the time evolution rule. A form using a model in which the operation state changes periodically or chaotically is positioned as an example of an embodiment of the present invention.

 (C-3) The first embodiment has been described assuming a system in which a large number of nodes distributed in space exchange data with each other wirelessly. However, the utilization form of the present invention is not limited to a system that performs wireless communication. It can also be applied to a system in which a large number of nodes distributed in space exchange data with each other by wire. For example, the present invention can also be applied to a wired LAN system such as Ethernet (registered trademark). Similarly, the present invention can be applied to a network in which different types of nodes are mixed, such as sensors, actuators, and servers connected in a wired manner. Of course, the present invention can be applied to a network in which nodes connected by wire and nodes connected wirelessly are mixed.

  Furthermore, the present invention can be used as a communication protocol for exchanging routing tables at different timings on the Internet. Here, the router is a relay device having a function of distributing the destination of information flowing on the network (communication path selection). The routing table is a communication route selection rule that is referred to when sorting information destinations. In order to realize efficient communication, it is necessary to update the routing table sequentially according to changes on the network, local traffic changes, and the like. For this reason, a large number of routers existing on the network exchange routing tables with each other at regular time intervals. However, in Reference 4 “Floyd, S., and Jacobson, V.,“ The Synchronization of Periodic Routing Messages ”, IEEE / ACM Transactions on Networking, Vol. 2 No. 2, pp. 122-136, April 1994.” As disclosed, it has been found that although each router transmits a routing table independently, a phenomenon occurs in which the routers gradually synchronize (collision). Document 4 proposes a method for dealing with this problem by giving random variability to the processing cycle of each node for a communication protocol used for exchanging routing tables, and a certain effect can be obtained. Is shown. However, since the method disclosed in the above document is basically a method that depends only on randomness, its effect is not sufficient.

  On the other hand, when the present invention is applied to the above problem, it is possible to autonomously mutually adjust the time slots for transmitting the routing table between neighboring routers. Accordingly, the transmissions of the routers have different timings, and a higher effect can be obtained compared to the method disclosed in the above-mentioned document 4.

  As described above, the present invention can be applied to the problem of collision and synchronization of outgoing data existing in any network regardless of wireless system or wired system, and is efficient data having both adaptability and stability. It can be used as a communication protocol for realizing communication.

(C-4) In the first and second embodiments, as an example, the natural angular frequency parameter ωi is assumed to be unified to the same value in the entire system. However, this is not essential for practicing the present invention. It is possible to operate even if the value of ωi is different for each node. For example, the operation can be performed even if the value of ωi of each node varies small according to a probability distribution such as a Gaussian distribution around the reference value.

(C-5) The present invention is characterized by acquisition control of communication timing information (phase signal in the embodiment), and it does not matter how the timing information is used for communication. For example, if the transmission frequency of the data signal from each node is different, communication may be performed without setting a time slot. Even in this case, the start of data communication can be determined from the communication timing information. It should be determined.

(C-6) As a patent application of a prior application related to the present invention, there is Japanese Patent Application No. 2003-328530, but the specification and drawings of the prior application patent application have numerous modified embodiments described therein. Among them, those applicable to the present application are also modified embodiments of the present application.

It is a block diagram which shows the detailed structure of the communication timing calculation means of 1st Embodiment. It is a block diagram which shows the node structure of 1st Embodiment. It is explanatory drawing (1) of the tuning between nodes in the communication system of 1st Embodiment. It is explanatory drawing (2) of the tuning between nodes in the communication system of 1st Embodiment. It is explanatory drawing of the phase response function R ((DELTA) (theta) ij (t)) which the communication timing calculation means of 1st Embodiment uses. It is a block diagram which shows the node structure of 2nd Embodiment. It is a block diagram which shows the detailed structure of the communication timing calculation means of 2nd Embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10, 30 ... Node, 11 ... Impulse signal reception means, 12 ... Communication timing calculation means, 13 ... Impulse signal transmission means, 14 ... Tuning determination means, 15 ... Data communication means, 16 ... Sensor, 21, 41 ... Phase calculation part , 22, 42 ... collision rate calculation unit, 23, 43 ... accumulated stress calculation unit, 24, 44 ... stress response function value calculation unit, 31 ... phase signal reception unit, 32 ... interference detection unit, 33 ... data communication unit, 34 ... communication timing calculation means, 35 ... phase signal transmission means.

Claims (11)

A communication control device equipped with a communication timing calculation means for determining the timing of data transmission from the own node based on the state of the phase that is internally changed, mounted on each of a plurality of nodes constituting the communication system,
The communication timing calculation means is
A phase calculation unit that changes the state of the phase of the own node according to a predetermined time development rule based on reception of the state variable signal from the neighboring node that reflects the phase representing the timing of data transmission of the neighboring node;
A collision rate calculation unit that observes a phase difference between the own node and a neighboring node and calculates a collision rate between the data transmission timing from the own node and the data transmission timing from the neighboring node based on the phase difference;
Stress response function value generation that accumulates stress values according to the collision rate in time, and causes a phase shift of random magnitude in the time evolution rule used by the phase calculation unit according to the accumulated stress values With a part ,
Continuously transmits a continuous phase signal as a state variable signal reflecting the phase of its own node, receives a phase signal transmitted by a neighboring node, and the above phase calculation unit uses the phase of the neighboring node. A communication control apparatus comprising phase signal communication means for enabling. The phase signal and the data signal communicate using different frequency bands, and the frequency band for phase signal transmission is subdivided into an integer number of bands,
Between the neighboring nodes, according to claim 1, characterized in that it comprises as an initial operation mutually exchanging the phase signal, a phase signal frequency band arbitration means each node arbitrates frequency band for the phase signal transmitter to use Communication control device.   The time evolution rule used by the phase calculation unit is that when the relative phase difference between the own node and the neighboring node is less than or equal to the phase width necessary for collision avoidance, the phase of the own node is 2. The communication control apparatus according to claim 1, wherein a phase response function having a mechanical characteristic that changes in a working direction is used. The said collision rate calculation part calculates the collision rate by observing the accumulated collision amount in a fixed time, and making the value normalized by the maximum accumulated collision amount into a collision rate . The communication control apparatus in any one . The stress response function value generating unit, the higher the collision rate, according to any one of claims 1 to 4, characterized in that to calculate the stress value based on a function having the characteristics shown a great stress value Communication control device. The stress response function value generating unit, a time interval for storing the previous claim, characterized in that up to the present time from the time of executing the random phase shift in response to the accumulated stress value 1-5 The communication control device according to any one of the above. The stress response function value generation unit evaluates an accumulated stress value at regular time intervals, and causes a random magnitude phase shift in a time evolution rule with a probability based on the evaluation value. the communication control device according to any of claims 1-6. The impulse signal as the state variable signal is discretely transmitted at a timing determined by the phase of the own node, the impulse signal transmitted by the neighboring node is received, and the phase calculation unit can use the phase of the neighboring node. the communication control device according to any one of claims 1 to 7, characterized in that it has an impulse signal communication means for. Node, characterized in that it comprises a communication control device according to any one of claims 1-8. Communication system, comprising a plurality distributed nodes according to claim 9. A communication control method including a communication timing calculation step for determining a timing of data transmission from the own node based on a phase state to be changed internally, which is executed by each of a plurality of nodes constituting the communication system,
The communication timing calculation process is as follows:
A phase calculation sub-step for changing the state of the phase of the own node according to a predetermined time development rule based on reception of the state variable signal from the neighboring node reflecting the phase representing the timing of data transmission of the neighboring node;
A collision rate calculation sub-process for observing the phase difference between the own node and the neighboring node, and calculating the collision rate between the data transmission timing from the own node and the data transmission timing from the neighboring node based on the phase difference;
Stress response function value that accumulates the stress value according to the collision rate in time and causes a random phase shift in the time evolution rule used by the phase calculation sub-process according to the accumulated stress value Generating sub-steps ,
Continuously transmit a continuous phase signal as a state variable signal reflecting the phase of its own node, receive a phase signal transmitted by a neighboring node, and the above-described phase calculation sub-process calculates the phase of the neighboring node. A communication control method comprising phase signal communication processing for enabling use.

Images